ESV.Eleventh.Int.Conf.SECTION.FOUR.Session.Six

Technical Session Six 

Heavy Duty Vehicle Safety 

Chairman; Dr. Lennart Strandberg, Sweden 

Large Truck Accident Exposure in the U.S. 

Hank Seiff, 

Motor Vehicle Manufacturers Association, 

United States 

Abstract 

In order to assess large truck safety, information is 

needed on vchicle exposure to accidents (travel) as 

well as on invo/vement in accidents. Since 1979, the 

University of Michigan Transportation Research Institute 

has collected detailed data on every fatal large 

truck accident in the U.S., about 5,000 cases a year. 

With the completion of the first comprehensive large 

truck exposure data survey, by UMTRI, we will be 

able to go beyond past analyses of large truck 

accidents. The National Truck Trip lnformation Survey 

(NTTIS) was initiated in late 1985. Data collection 

was completed in February 1987. For this survey the 

owners of over 4,000 large trucks were contacted tbur 

times over a twelve-month period (16,000 survey days) 

to obtain detailed information on the use of the truck 

on a randomly-selected survey date. The information 

collected includes the configuration, catgo, actual 

weight, and the route the truck followed. The combination 

of accident data with miles traveled from 

NTTIS will enable the calculation of fatal accident 

involvement rates by vehicle type, road class, etc. 

Background 

As the U.S. economy came out of the recession of 

1974 and '75, truck mileage increased. With the 

Table 1. Fatalitles & latality rates (u.S.). 

Year Combination 

Truck 

Fatalities 

1976 

1977 

1 978 

1979 

1980 

1 981 

1 98? 

1 983 

1 984 

1 985 

*preliminary 

3,909 

4,198 

4,643 

4,950 

4,238 

4,388 

3,911 

4,079 

4,257 

4,650. 

Combination 

Truck 

Mileage (x 108) 

57,937 

61,179 

65,636 

66,313 

67,386 

69,388 

71,129 

73,562 

76,986 

79,40? 

increase of mileage came an increase in fatal truck 

accidcnts. The subject of large truck accidents became 

a matter of increasing public concern. This concern 

has continued and intensified as the U.S. moved to 

economically deregulate motor carriers, with many 

claiming that lower frcight rates and increased competition 

led to decreased concerll with safety. Although 

table 1 shows that the increase in combination truck 

fatalities and fatality rate abated after 1979, it also 

shows that the combitration truck fatality rate continues 

to run about twice that of all vehicles in the U.S. 

Although manufacturers, carriers and governlnent 

saw the number of fatalities increase between 1976 

and 1979 they did not know the specific rcasons for 

the increase. Information on large truck accidents was 

fragmented, incomplete, and, in many cases, simply 

unavailable or so inaccurate as to be little value in 

determining accident causes. Yet accurate accident 

data are the key to finding effective solutions to truck 

accidents. Without goocl data, there is atr excellent 

chance that solutiolts based on intuition alone may be 

suggested. Such solutions may impose high costs 

without providing actual improvements. 

Truck manulacturers belicved that a comprehensive, 

coordinated research effort-involving government, 

industry, the academic and scientific communitieswas 

required to collect and attalyze in-depth data on 

large truck accidents. Such an effort would include 

collection of miles traveled (exposure) by different 

kinds of trucks, and could be used to identify the role 

Combination 

Truck 

Fatalities/108 

Miles 

6.75 

6.86 

7.07 

7.46 

6.29 

6.32 

5.50 

5.54 

5.53 

5.86 

All Vehicles 

Fatalities/108 

Miles 

3.25 

3_26 

3.26 

3.34 

3.34 

3.17 

2.76 

2.57 

2.58 

2.47 

;f 

'J

of vehicle designs, the effects of speed differentials, 

and other factors that had not been investigated in a 

systematic manner. 

The Motor Vehicle Manufacturers Association and 

the Westem Highway lnstitute took the initiative by 

sponsoring a comprehensive study of large truck 

accidents in 1979. The American Trucking Associations 

joined later. They spon$ored the University of 

Michigan Transportation Research Institute (UMTRI) 

in conducting this pioneering study and set the following 

goals: 

' 

r Determine the accident, injury and fatality 

rates (in terms of events per vehicle-rnile, 

ton-mile and/or cube-mile) lor a broad range 

of heavy trucks operating on U.S. highways. 

These should include at least comparisons 

among straight trucks, tractor-trailers, doubles, 

and triples; cabover vs. conventional 

designs; and, combinations of various 

lengths. 

r Determine the causes of accidents involving 

heavy trucks. 

r Achieve an understanding of the possible 

countermeasures which are likely to prevent 

or reduce the frequency of such accidents. 

The study is called "Acquisition/Analysis of Truck 

Accident and Exposure Inforrnation." Phase I (completed 

in 1979) assessed thc status of available data 

and determined what was needed to conduct a comprehensive 

analysis o[ medium and heavy truck accidents 

on a national scale. lt recognized the need for 

both accurate accident counts and mileage at comparable 

levels of detail if accident rflI€s were to be 

determined. lt pointed out the fallacy of comparisons 

which used simply "vehicle miles" but failed to 

recognize that, for example, a far greater percentage 

of truck miles than pa$senger car miles were accumulated 

on rural roads and at night. UMTRI concluded 

that "The availability and the nature of present 

. . . accident and exposure information . . . in much 

detail is rather limited." and outlined a program 

which would provide information to fill these gaps. 

Accident Data 

Phase Il of the UMTRI study evaluates all large 

(over 10,0fi) pounds gross vehicle weight rating) truck 

fatal accidents. lnitial information is taken from the 

National Highway Traffic Safety Administration's 

Fatal Accident Reporting System (FARS), and augmented 

with police reports and information from the 

Federal Highway Administration's detailed files on 

accidents in interstate commerce. When Federal Highway 

Administration accident reports are not available 

researchers go directly to vehicle owners, motor carriers, 

and drivers. This permits researchers to determine 

factors such as kind of road, weather, cargo, trip, 

638 

EXPERIMENTAL SAFETY VEHICLES 

weight, Iength, ownership, time of day and vehicle 

configuration. UMTRI publishes these analyses on an 

annual basis in the form of a factbook entitled 

"Trucks 

Involved in Fatal Accidents." 

These yearly factbooks, providing data on fatal 

accidents beginning in 1980, and special studies have 

already yielded insights, for cxample: 

Rollover is involved in almost 6090 of accrdents 

fatal to courbination vehicle drivers, 

ejection in 34V0. Extrication is involvecl in 

2290 (these are accidents in which the victim 

must be physically removed from the damaged 

vehicle; in some of these cases, the 

victim was crushed within the truck cab) and 

fire in 1690 (these figures do not add to 

10090 since more than one may be involved 

in a single fatal accident). 

The greatest number of fatal accidents take 

place on non-lnterstate rural roads (5490 of 

the total). Five-sixths of these are on twolane 

rural roads. Only 2590 of the fatal 

accidents take place on Interstate Highways, 

both urban and rural, yet combination trucks 

operate 4390 of their miles on the Interstates. 

A tractor-semitrailer is about three times as 

likely to have a fatal accident on a rural 

2-lane road as on an Interstate. Because 

vehicles are traveling at high speeds in opposite 

directions on two-lane rural roads. accidents 

can be more frequent and more serious 

than on Interstates or other divided multilane 

highways. 

Tractor-semitrailer combinations are involved 

in 72o/o of heavy truck fatal accidents, single 

unit vehicles in 2190, doubles or triples in 

390 and bobtailed tractors in 2Vo. 

Tractor-semitrailers have as many fatal accidents 

at dawn, dusk and night combined as 

they do in daylight. 

r Tank trailer combinations are about twice as 

prone to rollover as are van trailer combinations, 

r Rollovers are almost 5090 more likely in 

fatal accidents on dry than on wet pavements 

while jackknives are almost twice as likely on 

wet pavements. 

Data on all medium and heavy truck fatal accidents 

in the United States from 1980 through 1984 are now 

on the University of Michigitn's computer system 

undergoing analysis. Published data code books provide 

tabulations such as those illustrated in table 2 

below from "Trucks lnvolved in Fatal Accidents, 

1983": 

Data in the form shown provide basic information 

which can be analyzed to answer specific safety 

questions, for example the seriousness of the drunken

Table 2. Sample code book deta tabuletlone: 

"Trucks 

Involved ln Fatal Accldents, 1983"- 

variable ?07 

FREQ Prcnt 

4716 95.'l 

228 4.6 

0 0.0 

variable 2I0 

FREQ Prcnt, 

0 0.0 

132 2,7 

4477 90.5 

92 I.9 

r7 0,3 

19 0. 'l 

I 0.0 

2 0.0 

I 0.0 

203 4.r 

variable 1028 

FREQ Prcnt 

z62a 53.2 

2I95 44.4 

l2r 2.4 

DR,IVER DRINKIHG 

DRIVER DRINKING 

0. No drinking reported 

l. Drinking rePorLed 

9. UnknoHn 

LICENSE STATUS 

LICENSE STATUS 

0. None required 

l. Non6 

?, valid 

3, Suspended 

4. Revoked 

5. Expired 

6. Cancelled or denietl 

7. Learner's perrnrt 

8. Temporary 

9. unknovril 

CAB STYLE 

CAB STYLE 

SECTION 4. TECHNICAL SESSIONS 

I. conv6ntional 

2. Cabover or cab-forward 

9. Unknown 

driving problem among truck drivers, compared to the 

general driving population. 

In a study of truck crashworthiness now underway, 

UMTRI is examining the role of cab deformation in 

truck occupant fatalities and identifying specific ways 

in which occupants are injured in the cab. The 

discovery that 169o of truck occupant fatality accidents 

involved fire in somc way has led to in-depth 

investigations of accidents involving fire or fuel spillage 

to determine the role the truck fuel system plays 

in these accidents. 

It will be necessary to continue the accident data 

collection which has now been underway since 1980' 

so that both long ancl short term safety trends can be 

identified. This will also allow us to see the long term 

results of actions taken to improve the truck safety 

picture. 

Exposure 

Only a limited amount can be learned by studying 

fatal accidents alone. One needs to know something 

about accident frequency; how often do accidents of a 

certain type happcn in relation to vehicle miles travelerl; 

on what type of highways; and at what time of 

day or year. If accidents of a certain type are frequent 

there is a greater benefit to investigating them and 

finding a solution. And if a certain type of accident is 

found far more frequently on one type of road or 

during a particular time of day, a clue toward its 

cause may already exist' 

Phase Ill of UMTRI's overall study, called the 

National Truck Trip Information Survey (NTTIS) is 

providing information which allows accident frequency 

to be calculated. The truck trip survey provides 

information on truck population and exposure 

(miles travcled with detail comparable to that collected 

on fatal accidents). 

In the past students of truck accident exposure in 

the United States have gone to the Truck Inventory 

and Use Survey (TIUS), conducted every five years by 

the Bureau of the Census. Data for this survey are 

obtained by requiring a sampling of truck owners to 

fill out a form providing annual mileage and typical 

use of the vehicle. When the owner notes on the form 

that the truck is "typically" used in over-the-road 

service, other uses, such as mileage acsumulated in 

city dclivery, are not included. 

The NTTIS focuses on all mileage during a single 

day's use in order to get more accurate and detailed 

mileage data. The day's mileage is categorized according 

to road class (interstate, major artery, or other), 

rural versus urban, and day versus night' Urban areas 

are identified according to the FHWA definition. Two 

sizes of urban area are distinguished; areas with 

50,000 population or more' and areas with 5,000 to 

49,999 population. The NTTIS also categorizes milege 

by carrier operating authority, vehicle configuration 

(cab style, number of axles, trailer type and body)' 

length, cargo, cargo weight, and gross cornbination 

weight. This will provide information on truck use at 

a level of detail previously unavailable in the U'S. 

To provide a proper $ample of trucks for the 

survey, the University of Michigan worked with the 

R.L. Polk company, the only organization in the U'Swhich 

collects information on vehicle registrations 

ctirectly from the individual states. A stratified random 

sample of trucks registered as of July I' 1983' 

was chosen. The Bureau of the Census also uses R'L. 

Polk in setting up its sample for the Truck luventory 

and Use Survey. Based on both tunds available and a 

statistical evaluation of the level of data accuracy 

which might be expected a target sample size wa$ 

chosen. Owners of 2,000 tractors and 2,000 straight 

trucks were telephoned four times within the last year, 

providing 16,000 survey-days of truck exposure information. 

Since the operation of double trailer combinations 

is of concern, but still relatively rare in the 

U.S. the states of California and Michigan, where the 

majority of current doubles operation can be found, 

wefe oversamPled. 

It is recognized that the National Truck Trip 

Information Survey will leave many qucstions unanswered. 

There is the obvious statistical error in a 

sample of the size being used. UMTRI has calculated 

that we can expect almost a 990 error in average 

annual mileage for a category of vehicles comprising 

639

640 


one-fourth of our total, and a 20a/o error for a 

category comprising 590 of our total. For example 

since sales of low-bed trailers run about 590 of total 

Table 3. Estimates of the U.S. large truck (over 10,000 

lbs. GVWR) population*. 

trailer sales, we could expect to find a 200/o error in 

Sou!qq 

our calculation of their annual mileage. 

Truck type 

Beyond statistical error, the study is restricted to 

vehicles registered in 1983 while the survey of owners 

took place in 1985 and 1986. This was necessary 

Straight 

TractorE 

Truck 

2, 608, 100 

876,700 

2,068, {95 

886,643 

because good vehicle registration data runs that far 

TotaI 

3,505,00o ?,955.138 

behind the current year. So most 1984, 1985, and 

1986 model year vehicles are excluded from the 

rExcludlng Ala6ke, Hawaii and Oklahom! 

survey. Since later model vehicles are more likcly to 

be used for longer mileage, over-the-road operation, 

Table 4. Truck-trectors in use. 

and move to shorter mileage and local operation as 

they age, the survey should show lower annual mileages 

and a higher percentage of local and shorthaul 

in use I 

weekday 

49. rt 

I 

I 

t{eekend 

tO, Ar 

operation than is actually the case in the total U.S. 

fleet. 

not in uee I 50.9r | 89,?t 

Finally, and perhaps most importantly, as soon as 

the collected data is analyzed, many questions are sure 

Table 5. Stralght truck$ in use. 

to be found which these data do not answer. Already 

the questionnaire was modified to provide more 

information on rural versus urban operation of vehicles 

and the type of engines and fuel-saving equipment 

in use to answer questions raised by new U.S. 

ln use 

dot in us6 

I 

I 

weekday 

37,9r 

62. ft 

I 

I 

I 

Weekend 

B, lt 

rr. rt 

emission regulations. One must be reconciled to this 

sort of problem, knowing that good research ofterr 

asks more questions than could be expected when the 

project was begun. 

Although UMTRI has completed the clata collecrion 

phase of the National Truck Trip lnformation Survey, 

substantial time will be needed to complete the 

analysis of the data before it can begin to provide 

answers to many questions which will help us under, 

stand the nature of the U.S. truck fleet and pinpoint 

areas where safety improvements might be made. 

One of the earliest answers which ftas come from 

NTTIS is an accurate estimate of the actual population 

of large trucks (over 10,000 pounds gross vehicle 

The Future 

The combination of accident data with miles traveled 

from the NTTIS will enable the calculation of 

accident involvement rates by vehicle type, road class, 

etc. New and important insights into the causes, and 

hopefully, therefore the solutions to many large truck 

saf'ety problems, will be found. Yet it is easy to see 

there is much left to be completed. The various 

segments of the trucking industry, labor, insurance 

and government can all capitalize on the investment 

already made by providing sustained financial support 

for a comprehensive plan of action. Such a plan 

would include: 

weight rating) in the United States in the year 1982. 

Earlier population data, based on the 1977 TIUS and 

the Federal Highway Administration had been at 

variance*no one actually knew how many large 

trucks there were in the U.S.! Change$ were made in 

the TIUS calculation procedure between 1977 and 

I982 so there should be reasonable agreement between 

the NTTIS and TIUS numbers on nationwide truck 

population, since they came from surveys drawn from 

First The University of Michigan's fatal aoci. 

dent data collection program would be 

continued on an annual basis. lf it can 

be more closely allied with the Federal 

Highway Admini$ttation's accident reports 

and data collection than it is now. 

it can provide both the government and 

industry with information of better quality 

and completeness than it now has 

for analysis purposes. 

similar R.L. Polk samples. Table 3 shows the results. 

The NTTIS telephone survey has shown a relatively 

low proportion of trucks actually in use on any given 

day. This information must be tempered by remembering 

that the three latesr model year trucks (1984- 

86), those most likely to be in service more days and 

more miles, are not included in NTTIS as noted 

above. Survey data are shown in Tables 4 and S. 

Second The National Truck Trip Information 

Survey exposure study is collecting 

data for only one year. To meel future 

accident data needs, the survey should 

be continued on an annual basis. 

Again, such a program would supplement 

both governm6nt and industry 

knowledge and avoid duplication ol effort.

Third 

Fourth 

Fifth 

While the truck accident files created at 

the University of Michigan are the best 

available, they are limited to fatal accidents 

only. But fatal accidents represent 

just over one percent of all large 

truck accidents. The ovenvhelming majority 

of truck accident don't kill people, 

but they do injure people, damage property, 

snad traffic and sometimes, release 

hazardou$ materials. These accidents 

are probably as much responsible 

for the industry's poor safety reputation 

as are those in which people are killed. 

So it makes sense to collect and analyze 

data on the 99 percent of truck 

accidents that are non-fatal. Although 

NHTSA's National Accident Sampting 

System was originally conceived to provide 

this kind of data (for autos and 

trucks), it is being cut back substantially 

and will not provide the information 

needed on heavy truck safety. 

Analysis of the data already collected 

and of that which will be acouired in the 

future must be performed if the information 

is to be of use to decision-makers, 

safety officials, designers and operators. 

To date, very limited analyses 

have been made of these data. A great 

deal more can be done, because the 

information is now available, particularly 

for fatal truck accidents, in sufficient 

detail to isolate patterns of high frequency 

occurrences. 

The final element of the plan is the 

performance of special, in-depth, studies. 

One example of a special study 

which should be done concerns the 

accident experience of longer combination 

vehicles. The need for this study 

was pointed out in two Federal Highway 

Administration Studies of the benefits 

and problems of such vehicles. 

In-depth studies, such as those mentioned 

above on truck crashworthiness 

and fire and tuel spillage, can be used 

to pinpoint problems that are already 

identlfied, or those which show uo in 

accident studies. 


Having come this far, U.S. industry and government 

should now be prepared to takc the next $tep. 

The trucking industry is on the threshold of having a 

great resource at its disposal to help find solutions to 

truck accidents. But it will be wasted if we fail to 

continue to underwrite the costs necessary to continue 

collecting and analyzing accident data and making 

truck exposure surveys. 

It will take more than just one or two organizations 

to do the job. Government, insurance and academic 

interests should be involved in forming a coalition 

with manufacturers, suppliers, motor carriers, and 

other interested groups to provide sustained financing 

for this program. 

As I see it, the benefits are many. We shall have 

fact$ to counter growing criticism of the industry's 

safety record; the industry will have a solid basis for 

arguing for greater productivity in equipment and 

operations; and, it makes good sense to prepare now 

for truck safety issues that may become the subject of 

regulations in the future. 

Perhaps the greatest benefit-the one the public is 

most concerned about-is that we will finally have the 

information needcd to learn how to decrease the 

number of collisions between large trucks and passenger 

car$. That alone will make the whole effort 

worthwhile.


Vehicle Factors in Accidents Involving Medium and Heavy Trucks 

Robert M. Clarke and 

William A. Leasure Jr., 

National Highway Traffic Safety 

Administration, 

United States 

Abstract 

Among the many interrelated causes of truck accidents, 

vehicle-related topics play a critical, if sornewhat 

unrecognized and underreported role. ln many 

cases, these factors, if they do not directly cause an 

accident to occur, make it more difficult-or in some 

cases, impossible-for a driver to recover from an 

error or avoid an unforeseen conflict. Once a crash 

occurs, the way trucks are designed can affect the 

severity of the trauma sustained by the occupants of 

all the vehicles involved. 

This paper hiehlishts the fact that efforts to prevent 

truck accidents could be substantially aided by working 

to upgrade the performance of the truck brake 

sy$tems as well as truck handling and stability properties-especially 

as it relates to their tendency to 

rollover. Truck occupant crash protection could be 

enhanced by improving truck occupant restraint systems, 

providing a reasonable amount of protection 

from post-crash fires, making cab interiors free of 

sharp, hard objects that can cause injury during 

impact-especially steering wheel rims and hubs, and 

by improving cab designs to provide occupant survival 

space in a crash. Finally, an opportunity also existsby 

working on the designs of the front ends of trucks 

-to reduce the number of fatalities among occupants 

of other vehicles killed in collisions with medium and 

heavy trucks. 

Introduction 

This paper summarizes two recently completed Congressional 

reports (Sections 216 and 217 of the Motor 

Carrier Safety Act ot 1984, P.L. 98-554, October 30, 

1984)tlll21 on the general topic of medium/heavy 

truck safety. The two reports focus primarily on 

vehicle-related issues that influence the safety performance 

of medium and heavy trucks. These include: 

braking, handling and stability, crashworthiness, and 

truck occupant crash protection. 

Each report identifies the key issues related to each 

of these topics, summarizes what is known aborrt 

each, describes what might be done in the near terrn 

to make improvements, and lays out research agendas 

for the remaining longer term issues. 

The reports were developed with the assistance and 

participation of the complete range of interests associ- 

*Bracketed numbers indicate references given at the end of the papet. 

642 

ated with trucking, including: truck and trailer manufacturers; 

truck operators; driver groups; state, federal 

and foreign government research and regulatory 

organizations; representatives from the vehicle inspection 

ancl traffic law enforcement community; and, 

representatives of safety advocacy organizations. The 

Society of Automotive Engineers sponsored, in cooperation 

with the National Highway Traffic Safety 

Administration, a public symposium at which draft 

versions of the reports and the research plans were 

discussed and critiqued by all these interests' Consensu$ 

was reached that the reports had identified the key 

vehicle-related safety issues facing the trucking industry 

and that the research proposals contained in the 

reports reflect the best way of addressing those issues. 

Priorities for addressing these subject areas should 

be dictated by the size of the accident problem 

affected by each and the availability of achievable 

solutions. For this reason, efforts to improve truck 

brake systems should receive the highest priority since 

improvements achieved there are likely to be significant. 

Regardless of efforts to prioritize these topics, 

each plan represents a technically sound, consensus 

approach as to how each of the topics could be 

pursued, given that priorities and resources are allocated 

to that subject. 

The plans that are presented are not an exclusive 

agenda for government. Rather, it is hoped that they 

will serve as a blueprint that government and industry 

can use to address the topics discussed. 

U.S. Medium and Heavy Truck 

Accident Patterns* 

Heavy truck accident$ are complex, often lethal 

events that have many interrelated underlying causes. 

They include factors related to driver performance/behavior, 

vehicle pcrformance capability and condition, 

operating environment, and the amount and quality 

of safety management exercised by the motor carrier 

responsible for the driver and vehicle. Accidents occur 

when the "margin of safety" is reduced because the 

performance of one or more of these factors is low 

and compensation by the driver and,/or the vehicle 

cannot be or is not made. A balanced heavy truck 

safety improvement program, if it is to be effective, 

needs to be cognizant of the relationships among all 

these factors and must incorporate elements that 

simultaneously address all of them in some reasonable 

fashion. 

*lJnlcss otherwise noted, the accident statistics cited throughout this paper 

uere derived from: NHTSATS Fatal Accident Reporting System (FARS) and 

the National Accident Sampling Systcfi (NASS) for 1984, and tiom the 

States of Wa$hiilgl.on and Texas tbr l98l-1983.


Viewed comparatively, medium and heavy trucks 

are involved in a relatively small proportion of the 

Table 2. Occupatlonal fatelitiss-l984. 

overall number of motor vehicle accidents which 

occur each year (382,736 trucks were involved in 

lndustly CrauP Uorkers 

(x1000) 

Deaths* Deatha PEr 

100,000 Uork+r+ 

accidents in 1984-3.8 percent of the total). On the 

other hand, because of their size and a number of 

other factors, when they do become involved in 

AIl lndurtrlcr 

Trade 

Hirrufacturtnt 

servlc:+ 

covernnent 

Transportatlon 6 

104,300 

24,000 

IS,000 

28,900 

15 ,90o 

1r, 5oo 

1.200 

1,100 

1,200 

I, t00 

II 

5 

6 

9 

accidents, they are often severe. 

Pul,llc Utllltlss 

Constructlon 

5,500 

5,700 

1,500 

?,2OO 

2l 

As a result, their proportional involvement in fatal 

Asrlculrure 3,q00 1,600 

46 

accidents is higher (5,188 trucks were involved in fatal 

Tl{ucK DRrvERs I,876** I,087*+* 

58 

accidents in 1984-8.9 percent of the total). Normaliz- 

Htntn6. QusrrylnB l,0oo 600 

60 

ing for exposure, heavy trucks experience less non- 

SOURCES: *Ac.: I dent FFeG-l-gg-E, Nst ional Saf e ty counc rl 

**Edul.lv,nerrt snd Xffitbrp* JsnuuJ-I9-8-l, U. S . Deparimcni df: 

fatal accidents per lfi) million miles of travel than do 

passenger cars (288 versus 614, in 1984), but experience 

more fatal accident involvements per 100 million 

Labor 

miles of travel (4.0 versus 2.8, in 1984) than passenger 

cars. 

side fixed objects) than are passenger cars and light 

One way of gauging the relative importance of trucks/vans. 

medium and heavy truck safety is to assess the Medium and heavy trucks, like most other vehicles, 

consequences of these vehicles' accidents in terms of experience most of their accidents on the roadway 

the total number of fatalities and injuries that result. 

Viewed in this way, medium and heavy truck acci- 

itself (79 percent)-as opposed to off-road, in daylight 

(76 percent), on straight (79 percent), dry (66 percent), 

dents result in 12.8 percent (5,657) of all highway and level (69 percent) roads. All vehicle types have 

related fatalities and 4.8 percent (171,232) of the 

injuries that occur in highway related accidents each 

similar patterns. Observed variations from $tate to 

$tate are more indicative of geographic or weather 

year. The majority of these (118,835 of the injuries pattern differences than they are of differences in 

and 4,019 of the fataliti€s) were sustained by occu- truck accident involvement propensity. 

pants of other vehicles involved in collisions with 

medium and heavy trucks (see Table l). 

Truck drivers are involved in one of the nation's 

most hazardous occupations. They sustain 9'3 percent 

of all work-related fatalities, yet comprise only l'8 

percent of the employed work force. Truck driving 

rank$ second only to mining and quarrying in terms 

of occupatiottal fatalities that are sustained per 

100,000 workers per year (see Table 2). 

If a medium or heavy truck is involved in an 

accident it is most likely to be a collision with another 

motor vehicle. This pattern is typical for most other 

vehicles as well. Trucks are, however, proportionally 

more involved in single-vehicle accidents (rollover, 

loss-of-control/jackknifes, and collisions with road- 

Combination-unit trucks experience a large proportion 

(59 percent, Washington l98l-83) of their accidents 

on roadway types likely to be used in over-the' 

road operations (i.e., lnterstates, U.S. and State 

routes), This conttasts with the accident experiences 

of single-unit heavy trucks which reflect travel patterns 

in urban/suburban settings (68 percent of singleunit 

truck accidents occut on city streets or county 

roads). The speed involved with travel on the former 

types of roads, has a direct effect on accident severity 

outcomes. 

A large portion of combination-unit truck accidents 

(55.8 percent in Washington) occur on undivided 

highways. Travel on undivided highways provides an 

increased opportunity for the truck to be in conflict 

with other vehicles. Also, the occurrence of an acci- 

Table 1. Gonaequenceg of medlum gnd haaw truck 

accldents in 1984. 

dent on this type of road increases the likelihood of it 

being serious, since head-on collisions are possible' 

As previously discussed, there are typically numer- 

ttedih and Hr.5, Truck Occupsntr 

OccupsntE of other VehtcIBs 

Involved in cal.ll!tlons with 

Hedlun ilnd Hesw Tru.kE 

Pedesrrlans/cycl I sts InvoIv€d 

Hl I Ied 

1,087 

1,019 

rr8,815 

ous overlapping factors which combine to ultimately 

"cause" an accident to occur. Some of these are 

documented in accident data collection systems' 

Driver-related errors, infractions, or misjudgments are 

ln Accidenrs sirh 

H+svy Trucks 

flsdlM Bnd 

5sl 9,398 among the frequently cited factors contributing to the 

TotsI 

s,657 171.21? cause of truck, as well as other types of vehicles', 

loral (all htSlp.y frhr.d 

44.24t 3 . 573 ,2 Io accidents. In Washington, for example, in 54 percent 

12.8r of All FatalirtsE 

of all accidents in which combination-unit trucks were 

SoUBCES: Fes 1984 tnd tlAss 1984 

involved, the truck driver was cited for some type of 

error or infraction. 

643

While significant reductions have been made in 

recent years, alcohol is still involved in 43.3 percent of 

all fatal accidents. Alcohol is not involved proportionally 

in as many medium and heavy truck accidents, 

however. either in terms of the truck drivers or the 

other vehicle drivers involved. Based on a 15 state 

sample of fatal accidents where blood alcohol concentration 

(BAC) levels of fatal accident involved drivers 

are routinely gathered, it was found that 2.9 percent 

of all truck drivers, and 16.6 percent of the drivers of 

other vehicles involved in accidents with heavy trucks, 

had BAC's greater than 0.1. 

Factors related to the mechanical condition of the 

truck are sometimes noted as having contributed to 

the cause of an accident. Problems of this type are 

typically coded in most accident reporting systems 

only when equipment is obviously broken or worn 

out, as determined by visual inspection. Equipment 

that is degraded, but still intact, such as brakes that 

are out of adjustment is usually not reported. For 

example, in Washington in 1981-1983, only 8.9 percent 

of all the combination-unit truck accidents were 

cited as being attributable to vehicle component part 

deficiencies. Brake system deficiencics are the most 

prevalent. This contrasts with roadside vehicle inspection 

findings[3] where, routinely, 20 percent or more 

of the vehicles inspected are placed out-ol'-scrvice for 

vehicle component part deficiencies, most of these 

being related to brake system deficiencies. 

ln summary, medium and heavy truck accidents are 

not particularly numerous nor are they ovcrrepresented 

among all motor vehicle accidents. They are, 

however, unusually lethal and more often than not, it 

is other highway users, with whom trucks share the 

highways, that are the victims in these accidents. 

Among the most significant reasons why this pattern 

of fatal accidents occurs are: the large disparity in size 

and weight between trucks and other vehicles, the 

typically high travel speeds at which trucks are operated 

and, travel patterns that, in many cases, place 

them on undivided highways where the likelihood of 

collisions with other vehicles increases. 

Medium and Heavy Truck Dynamic 

Performance 

As with all motor vehicles, driver control of medium/heavy 

trucks is limited to braking, acceleration 

and steering inputs. Any or all of these control 

applications are utilized to operate the vehicle under 

routine conditions or in the attempt of non-routine, 

often severe, avoidance maneuvers when the driver is 

confronted with a potential crash threat. ln the case 

of most four-wheel vehicles, comparatively severe 

levels of either steering or braking must be made to 

induce dynamic instabilities in the vehicle. This is not 

the case with medium,/hcavv trucks. These vehicles are 

644 


susceptible to rollover, spin-out and jackknife in 

much less severe steering maneuvers. 

The ultimate criteria for judging the stability and 

control performance of a motor vehicle is whether or 

not the vehicle's driver can maintain stable control 

under all intended and foreseeable conditions of 

operation. In this regard, one can consider that the 

expectation of good dynamic behavior is fulfilled 

when the vehicle: 

I Attains a desired deceleration level during 

braking, 

I Follows a desired path in response to steer- 

I Remains upright (i.e., does not roll over), 

I Maintains a limited swept path, and 

I Does not oscillate from side to side 

uncontrollable manner. 

an 

In practice, niediunrlheavy vehicles often fail to 

meet these desired criteria for a variety of reasons. 

For example, they have the following performance 

limitations: 

Poor wheels-unlocked stopping performance. 

This results primarily from the general mismatch 

between the brake torques developed 

at each wheel and the prevailing wheel loads. 

This mismatch occurs due to the trernendous 

changes in wheel loading (both static and 

dynamic) that take place as a result of 

payload weight and placement. In addition, 

truck brakes often fail to deliver their designed 

torque output because they are not 

properly adjusted. 

Poor retention of braking capacity during 

desrcnt o.f long and/or sleep grades. The 

braking horsepower necessary for a fullyloaded 

vehicle to safely descend a substantial 

grade at highway speed places a large demand 

on the capacity of most truck brake 

systems. Parasitic losses which would normally 

aid in slowing the vehicle are low 

relative to the total vehicle weight. The 

search for improved fuel economy continues 

to reduce these parasitic losses even furthcr, 

Loss o! directional conlrol. Exceeding the 

vehicle's yaw stability lirnit results in vehicle 

spin-out (single-unit trucks), and jackknifing 

or trailer swing (combination-unit trucks) 

conditions. The primary cause oF these phenornena 

is the rearward bias of braking 

forces typical in the brake system designs of 

U.S. medium,/heavy trucks. This increases 

the probability of rear lvheel lockup. When 

lockup occurs, tires lose their ability to 

generate side force and the vehicle becomes 

unstable in yaw. Unstable yaw response in a

medium/heavy truck is likely to generate 

turning respon$es which exceed the vehicle's 

roll stability limit, thus precipitating a rollover, 

"Crack-the 

whip" response of multiplyarticuluted 

vehicles (doubles, triples and certain 

tntck-full-trailer trailer combinations). 

Multiple-articulated vehicles, have a tendency 

for the rearmost unit of the vehicle to show 

exaggerated or amplified response relative to 

the towing unit in certain types of severe 

obstacle avoidance maneuvers, 

"Rearward 

amplification" has important safety consequences 

when, during such maneuvers, the 

rearmost trailing unit exceeds its own roll 

stability threshold and rolls over. 

Straightforwurd vehicle rollover. Attempting 

turning maneuvers at too high a speed results 

in the vehicle's roll stability limits being 

exceeded. 

The stability and control characteristics of medium/ 

heavy trucks are direct indicators of their safety 

performance. A driver's ability to control his vehicle 

is ultimately limited by the response of the vehicle to 

steering and braking inputs. Limitations on the dynamic 

control capabilities of the vehicle reduce the 

viable options which are open to a driver in maneuvering 

to avoid traffic conflicts produced by other 

vehicles and also reduce the tolerance which is avail- 

able to compensate for any inappropriate control 

inputs made by the driver. In effect, the vehicle 

becomes less forgiving of control errors. 

The Performflnce Characteristics of 

Medium and Heavy Trucks in 

Maneuvers Involving Braking 

The Size Of The Brake System Related 

Safety Problem 

There are basically four different types of truck 

accidents that could be related to braking system 

performance: accidents due to failed or inoperative 

brakes; runaways on down grades; accidents where 

the vehicle was unable to $top in time (brakes did not 

fail nor were they ineffective due to heat but they 

simply did not provide the stopping force necessary to 

avoid the accident), and; skidding or loss-of-control 

accidents where wheels locked during braking. 

Collectively, the performance of truck brake systems 

could be a contributing factor in as many as 

one-third of all truck accidentsIl]. 

Truck Brake System Limitations 

Truck brake systerns have a number of critical 

limitations, namely: 


Inadequate Capacity in Continuous or Repeated 

Braking Situations-The adequate sizing 

of truck brake systems in terms of 

braking torque and thermal capacity is dictated 

by more than just the mass of the 

vehicle. For example, a tractor-trailer typically 

weighs approximately 30 times as lnuch 

as a passengcr car but needs 167 times as 

much braking power to maintain a steady 

speed on a 6 percent grade[4J. In addition, 

the capacity of truck brake systems has not 

increased to match the increasing demand 

placed on thcm as a result of fuel ecorlomy 

enhancement efforts to decrease parasitic 

drag. As a result truck drivers must compensate 

even rnorc than in the past, especially 

when descending grades. Lower descent 

speeds (and lower transmission gear ranges) 

must be used to prevent runaways. 

Poor Brake Distribution-U.S. trucks and 

combination-units typically have a strong 

rearward bias in the application of braking 

force. Front wheel/steering axle braking is 

usually low. This results in stopping distances 

which are longer than those of other 

vehicles, especially under emergency conditions. 

Additionally, combination-unit trucks 

can easily become unstable due to locked 

wheels under many brake application conditions. 

lncompatibility of Tractor and Trailer Brake 

Systcms-Many tractor and trailer brake systems 

are not compatible-i.e., they do not 

function well together to provide desirable 

overall combination-unit vehicle braking performance. 

Often, the atnount of braking 

force being applied by the tractor's axles 

greatly exceeds that of the trailer's, or vice 

versa. Similarly, the brakes may apply or 

"come 

on" quicker on the tractor than on 

the trailer. lncompatibility cornpromises both 

vehicle stability and brake effectiveness 

which can result in uneven brake wear problems 

and brake fade on downhill descents. In 

addition, brakes on trailers often apply and 

release slowly compared to those on the 

tractor. This is due to the distance between 

the brake control valve (treadle) and the 

trailer brake valve(s). Slow brake application 

times increase $topping distance and slow 

release times make it difficult to recover 

quickly from trailer wheel lock-up should 

this occur. This problern is more pronounced 

with longer combinations. 

Sensitivity to Brake Maintenance*Because 

truck brake systems are more complex and 

645

experience comparatively more severe service 

conditions than passenger car brake systems, 

they require a great deal more maintenance. 

Frequent inspections and repairs must be 

made to assure that systems are operating 

and are properly adjusted (since, unlike passenger 

car brake system$r most truck systems 

do not self-adjust with wear). Roadside in- 

' spections have, for many years, indicated 

that many operators do not adequately maintain 

their vehicles. This is compounded by 

the absence of consensus measurement standards, 

performance criteria and marking,/labeling 

schemes for components within the 

brake system. This makes it difficult, if not 

impossible, for truck operators to obtain 

replacement parts such as valves and linings 

which exhibit comparable performance to the 

parts that were originally installed on the 

vehicle. Because of this, compatibility prob- 

, lems are often created or worsened when 

repairs are made. 

In order to address these problems, a three phase 

program of research is suggested. It would deal: first, 

with compatibility and brake maintenance problems; 

secondly, with controllability problcms associated with 

braking maneuvers, especially while operating liehtly 

loaded or empty on slippery road surfaces, and; 

thirdly, with eftorts to optimize the brake system to 

improve stopping performance. 

Research Program To Address Brake 

Compatibility, Brake Adjustment, And 

Component Performance Problems 

There exists a need to ensure that today's brake 

systems function as well as possible. In any discussion 

of heavy truck braking systems-especially among 

truck users-the subjects of compatibility, component 

Ievel performance, and brake adjustment always surface 

as priority concerns. The reason for this is that 

poor compatibility becomes obvious very quickly in 

terms of excessive brake lining and drum wear, brake 

drum cracking, the need to adjust brakes con$tantly, 

etc., on the "over-braked" unit of the combination. 

The safety implications of operating a truck with 

incompatible brakes are subtle and difficult to i$olate 

in accident statistics. Marginal stopping perlbrmance 

would not, of itself, precipitate an ascident. lt only 

becomes a problem il a crash avoidance braking 

maneuver is attempted, and these are nol cvcryday 

occurrences. In addition, if a crash does occur under 

these circumstances, other more apparent factors are 

likely to draw attention away from the fact that poor 

braking performance was a contributing factor. 

The overall research and implementation program 

needed to address brake system incompatibility and 

646 


brake maintenance issues is shown in Figure l. 

Industry, government and professional standards setting 

organizations are currently actively involved in 

programs to address the issues of pneumatic timing, 

brake force balance, performance and labeling of 

valves and linings and improved means for ensuring 

proper brake adjustment. 

These activities need to be actively supported and 

continued, and, where possible, accelerated. Compatibility 

and maintenance issues are the everyday concerns 

of conscientious motor carriers. They must be 

satisfactorily addressed before motor carriers will 

become more receptive to the new technology that 

must be incorporated in trucks in order to make 

significant improvements in truck controllability and 

stopping capability when braking. 

Research Program To Address 

Braking-Induced Instability Problems 

Even if substantial improvement can be achieved 

with regard to tractor-trailer compatibility and maintenancc 

issues, truck brake system performance would 

still be deficient at limit conditions, i.e., when making 

an emergency stop or when making a brake application 

that is too "hard" 

for conditions. An example of 

the latter case would be when the driver has misjudged 

the amount of brake pressure he can sal'ely 

apply when operating an empty or lightly loaded 

vehicle on a slippery roadway. Without load proportioning 

systems and/or antilock braking systems, compatibility 

can only be achieved lor a single design 

loading condition, typically the fully loaded condition. 

Many truchs operate lightly loaded or empry a significant 

portion of the time. 

The most promising technology that is currently 

available for significantly improving braking performance 

at these limit conditions is the antilock brake 

system (ABS). These systems are the only solution to 

the wheel lock and resultant loss-of-control tendency 

typical with currently designed U.S. vehicles. Almost 

everyone in the trucking industry agrees that antilock 

has the potential to significantly improve the braking 

performance of heavy trucks by eliminating the directional 

instabilities which occur when wheels lock. 

Many, however, question the reliability and maintainability 

of the $ystems in actual use a$ well as the 

ability of the systems to fail safe (i.e., in the event of 

a malfunction the $ystem reverts back to a normal 

brake systcm without antilock). Lack of reliability was 

the major reason for the C

tr.cr0r/tr.ll.r coilt.l'blr'it 

t D.rrloF l.rl prec.dur...nd 

p.rlornrnr.'.n'd0p.' 

a F..6iv. cafiiltnla lo 

a r.r!. lrn.l rul. 


Cpilpoornl L.v.l F.rlot-.4t 

I D.v.loO fi.r.vr.hr^l 

Er 0(.4!r.. .id F.rlornrnc. 

I En.ur. cone.rbllilr ol 

P.tr. tlti lh.l 

,i'Eh ll r.or.cr. 

a E.r.bll.n ao.lr ror Inl{rlrt 

Prad!.1 

lr.l. Y.lr.. 

a Ar..rr p.rlorfi.ficr ol 

r.pr.r.nr.t'r. rrlrrr 

-Crr.r prarrsrr 

-irlr.r. 

Dr.r.s.rr 

d.r.locrrnl 

a P!!rrcrrrDn ol t..t/ 

I Inlor4.r'on/.rt.rlr.c. 

0rrl. Ltfrtfitr 

I F...{rr iqrqs. rr I 

-Fr.ror6.n€. rattnl 

O.Yrroq Fr,ror6.ncr Crrtrrlr ro 

E..ur. Optriluh Frrlqrr.nc. ot 

Curr..r ft.r.8tttan. 

l{elpil.rt/Qortofi..l Flt, 

Co...n.ur 9tr6d.rdt/ 

F..oFir6{.d Fr.cilcrr 

a F.lo.6rnt. r.rl 

t E.rr tnj tt.b.t ..nt tâ 

}rl. Adtrthtrl 

I Drvrtop pr.lorr.ncr t..t 

a Erl.bll.fi rcr.pr.ot. 

t.rlorfi.n.. rrilri. 

a ldr.tllr rr.0.rr lor lidlttrt 

d.r.lqFi.nr ol rp9..d.d 

Figure 1. Truck brake performance improvement program-brake compatibility and malntenance 

closely monitored fleet study, in cooperation with 

antilock suppliers, truck manufacturers and motor 

carriers is suggested. This study would yield sufficient 

performance, reliability, maintainability and cost data 

to support intelligent decision-making on the part of 

motor carriers and governrnent relative to the suitability 

of this technology for widespread application in 

trucking, Initiating such a study now would result in 

the data being available in the early 1990's. Thus, it is 

imperative that this portion of the brake rcsearch 

program be conducted in parallel with the cotnpatibility 

research project discussed previously. The project 

is outlined in Figure 2. 

Research Program To Improve Truck 

Stopping Performance 

The objective of this part of the program would be 

to achieve the maximum practical limit braking performance 

possible. It would build on the improvements 

expected to result from the first two portions of 

the program 

Ultimately, a vehicle's stopping performance is 

limited by the overall amount of brake force capacity 

that foundation brakes can generate and by the 

traction properties of the vehicle's tires. Truck brake 

Ost-Or-.dlvrt'..r 

r.drc.ro.r 

a Erl.btr.n il.lhad. cl 

I qetlrct rhnrri rrDftG.t. 

Pfoltlrt. Oroducr. 

I Flrll.rrlr.l. 

force capacity on domestic vehicles is already at its 

limit with the exception of front wheel brakes. The 

power of this part of the system could be increased as 

could tire longitudinal traction. Research is nece$sary 

to understand the trade-offs involved in achieving 

these objectives and to establish reasonable goals for 

product development by the industry. 

This part of the overall program would, in general, 

follow the previously discussed research. It is shown 

diagrammatically in Figurc 3. 

The Performance Characteristics of 

Medium and Heavy Trucks in 

Maneuvers Involving Steering 

The steering responsc characteristics of a vehicle are 

one of the principal descriptors of its safety performance 

capabilities. In addition to braking capabilities, 

these properties define the inherent limits of safe 

vehicle operation. 

As a result of extensive research conducted since the 

early 1970's, considerable progress has been made in 

identifying the factors which affect the dilectional 

control and stability of medium/heavy trucks. The 

current state-of-knowledge indicates that some trucks 

64'l

Lr.g.-Scrl. ln-$.rvlcr Ou{nllllcrllon 

of Accldifrt Roduqllon B.n.llt 

F.d€.il vohlcli/Eqqlgmonl sl.ndirdr 

i ForlorFtnca sDeclllcallq4r for n{* 

Yohlcla ryatamr 

Urer/Oo.rdlqr Slrndrrd., 

Rrcomnrnd.d Prrdllcol 

I Comprrlbtllty ('mrrlng') sdldrllnar fol 

lrrctorr and lrrlloaa vir-r-ylr rntllocl 

rnd/o. lord p/oForllqnlng lyrlrml 

I Dflv..rncchsnlc ar.lnlng.ldt 

t DrlY.r tt.rlogy llqld.llnr. 


Anltlorh lftlllrltYi 

I Tott t.ACk oYsludllonr 

-Perlornrnca of cuarenl rnlllocl 

dealgnE 

-Trsdoofls b.trdsn control rl.ttoglal 

trnd pcalo.monco 

-Enelneerlno anqlyflg ol dur.blllly 

-Cohp.tlblllly 

ol combldrllqnr 

(q{rlorFanca slth rnd wllhqul 

rfrlllooh ) 

l^-$ervlc. Fl.et Tdtll 

r F'llrblllty/du.rblllly 

t Faltura fiodg arrerafrint 

a ll.lnl.nanc. r.mlf lcrllonl 

r Paafornrnc. FtrlqrmancarRelliblllty D.ll 

Orclrlon on N.id rqr Srfoly Po.lorilrnE. 

Flict Evrlq.llon. 

F.d.nl Slrli Iniprcllon/M4lnltnrnc. 

Slrndrrdr 

I CompllFcntrry lo n.s yihlqlr rid. 

ll lrrsrd 

a EUTOOO/ Aqrtr.llr 

aU.S. motor ca.rl.. ivtlutllonl 

I Publlcti{on sf trnt rcrsllt .nd proortrl 

. D.mOqtlrrtlont/f llmr 

a Fubllc dlrcurrlon of .itqlla 

a Drtvrr lr.lnlog .ldl, drlvl. ltalliqt 

O{ld.llm. 

Figure 2. Truck brake performance improvement program-braking-lnduced Instability 

have safety-related response tendencies that limit the 

range of performance over which they can be operated. 

Heavy trucks cannot be steered around corners, 

change lanes, or avoid unexpected obstacles as quickly 

as a car can, nor are they able to make right-angle 

turns the same way cars can without experiencing 

difficultics. They are more prone than cars to rolling 

over in turns or, in the case of some multiplyarticulated 

combination-unit trucks, when attempting 

quick lane change accident avoidance type maneuvers. 

Multiply-articulated vehicles also may exhibit oscillatory 

behavior whetr simply travelling in a straight line. 

Of the topics previously discussed, rollover has 

direct and significant safety consequences and is in 

need of additional work to translate previous research 

findings into implementable solutions. Accordingly, a 

program for further re$earch in this area is proposed. 

Rearward amplification is a problem unique to a 

special class of vehicles (multiply-articulated, larger 

combination unit vehicles (LCV's)). Some vehicle 

design-related changes could be made that would help 

reduce the likelihood of this occurring. These are 

close to being implementable. Other factors also 

64E 

Eqolpmrnl/Cohponrnl Llg, Conr.nru. 

StdrrHaco|rlsrf, drd Procrdutrr 

I Cofr$onint tnd lull tyrlem levtl 

psrlornrnci lill frorturomtnl 

Drociduail 

-Mi.ilnO/lrb.llng gsldrllnar 

affect this tendency, and, in many cases, to a greater 

degree than do vehicle-related factors. These include 

operational use practices and legislative choices relating 

to vehicle size, weight, and configuration allowances. 

Problems associated with low speed off-tracking are 

certainly a concern from a traffic cngineering and 

operations viewpoint, but are not significant from an 

highway safety viewpoint. Few traffic ascidents are 

likely to be associated with this characteristic of 

trucks. Those that do occur are likely to be low 

severity, property-damage-only events. Accordingly, 

no additional work on this subject is proposed. 

High-speed yaw instability, while demonstrable 

from an engineering viewpoint, is not evident a$ an 

accident causal factor. lt is not likely to ever be 

evident in mass accident data files. since when it does 

occur, it is likely to rcsult in the vehicle rolling over. 

This topic is best addressed in conjunction with 

efforts to improve truck roll stability. 

The oscillatory behavior of multiply-articulated vehicles 

is also not likely to be a signilicant highway 

safety problem. It is of concefn, howevet, in that 

trailing units may encroach on other travel lanes or

Alt€thstrvo,rldnorEtivs B.Eko 068ign 

Evalvrtiona (TrEt frdck) 

a Eydluot6 aoloty psrlo.manc. grlfif ol: 

-Lord-senrilivt Dtlta Prooorliontng 

-W{dOa/dirc 

-Elactronlc 

rctlyrllon 

-Olha,. na* tyttamt In loffrt ol: 

--rblllly 

lo mrlnrrln adjuEimenl 

-'unllo.m torqua output ovrr taaylca 

il10 

Frdatrl Eoulpm.ftl Slrndirdt 

a Upgrrd. sy.lafi Darlormancl 

rrqglrrmtnla rr lpproprltli 


lftcroasod Flont Whoel Brsho Forcr 

Cagrcily (T.Et TrEck) 

aABtaBt po.lo.manc. grrn tttoclrlod filh 

biO0e. b.rkes 

aEYaluato aamltldttlons odi 

Slaorlno Ind rurp€nsion tyttimr Ind 

dtlvor rccaplrblllty 

allrarure rteaalng .itDonta chalacttllallcl 

Ol Yrhlclrt wilh: 

-lncrttalnq ailounla ol b.rke lorqur 

*vrri6ul lygat ol tlsg,ing rldf (e.0otf..t 

klnOpin ry.teor, pswtt rlFringl 

atletina teri morauasfront oaocgdqao lot 

t!ttbllthlng ateering reNpOore 

chrarclarlallcr 

Flrol Eyrlsallon ol Paomlrlno lrcnnology 

, 

a Patlotmtnci 

rF.il.btilfy 

aM.tnt.tnrbilttt 

lCort 

Equlpm.nirCompdnanl MIO- Conrintur 

Sltddsrdr 

t Dav.log tlr. trrctlqn otttu.iminl 

Daoctduae 

a DaYrlop tractlotr ratlnO .cheD. 

I Erirbllrh perforfitnr6 goslt loi 

product devolopmanl by lha lndualty 

Figure 3. Truck brake performance lmprovement program-stopping performance 

intimidate other motorists and thereby cause erratic 

maneuvering 

actions. 

The Prevalence And Characteristies Of 

Rollover Accidents 

Rollovers constitute a vefy visible and serious type 

of comrnercial vehicle craslr. Although vehicle rollover 

is involved in l'rom 4 to g percent of all medium/ 

heavy truck crashes, it accounts for approximately 

one third of the single-vehicle accidents. Rollover 

occurs in approximately l5 percent of the f atal 

crashes and is a contributory factor in nearly 60 

percent of the medium/heavy truck occupant fatalities. 

Research Plan For Improving Truck 

Handling And Stahility Performance 

Rollover is given the highest priority among handling 

and stability related issues because it is well 

undcrstood and has an obvious direct link to safety. 

The program to improve the roll stability properties 

of trucks follows three parallel paths. 

Tire Traction (T€si Track/L.b) 

aAEs€aa frnga ol tonOitudinrt ind 

l6l6rdl lr6ction crOrbrilty of 

lodty'a il.et 

aEvrlurta tlrdiollr baiwaan 

-Trrctlon 

-tryaar 

-Cort 

-Ou..blllly 

a Pdrrmotrlc liEia 

-EEltblisi porlormtnca trhga 

tt t luncllon ol: 

-Comgoundlno/brrnd 

- wia. 

-Trrrd Drilirn, atc. 

-Divolop oerr0ramaftl 

melhodolo0y 

lnlo.mrtlon Dlrdemination 

I Publlrh t.tt llndlngr 

a F.brlcrle r hybtld 

rlrlo-ol-the-ral Yahlclc tnd 

demonrtrrl. lli caPibllllY 

a PsbllBh rrlltrgr 

One of the paths would be directed towards developing 

the best methods of gauging the relative roll 

stability perforrnance of trucks. It would take into 

account static and dynamic considerations. 

A second path would attempt to establish what 

motion and visual cues drivers sense (or possibly fail 

to sense) prior to a rollover. This information could 

help driver training efforts and could possibly result 

in more of the "good" cues being built into future 

trucks. 

Another path would study in-service trucks to a$sess 

how many of them are typically being operated close 

to their stability limits. This would include studies of 

truck tires to determine the degree to which their 

performance propertie s affect vehicle stability and 

control. A determination would also be nrade as to 

which properties of in-service trucks are most responsible 

for stability limits being approached. Finally, an 

assessment would be made of the impacts that would 

result l'rom design-related cltanges that might be 

contemplated to enhance roll stability of future 

trucks. 

649

The overall roll stability enhancement research program 

is shown in Figure 4. 

Medium and Heavy Truck Crash 

Performance-Truck Aggressivity 

When any two vehicles of dissimilar size collide 

with each other, the larger of the two vehicles 

typically inflicts much more damage and injury 

trauma than it sustains. In this case, the larger vehicle 

is said to be more *'aggressive" relative to the smaller 

vehicle. 

In medium and heavy truck collisions with other 

vehicle types, the truck's "aggressivity" occurs for 

two principal reasons: geometric mismatch (the fact 

that the physical shapes of the two vehicles, particularly 

the front end of trucks, do not match each 

other), and mass mismatch (the difference in weight 

between the two vehicles). 

Many believe that little can be done about truck 

aggressivity short of segregating trucks from other 

vehicles. This may not be totally true since a significant 

portion of the problem arises from geometric 

rather than mass differentials. Practical improvements 

may be possible for at least this part of the problem. 

Extent Of The Aggressivity Issue 

Two-vehicle collisions are the largest single category 

of fatality producing motor vehicle,/highway related 


Flgure 4. Truclr roll stability enhancement research program 

650 

Tarl Procadurt OQY6lopfienl 

aDtlrrmlnr tlrllc mtllurimanl 

Irchnlquar' rDllllY lo ritoir 

dynrmlc aflecrt 

tTrrl Prlcadura DtY!loPfttnt 

-PowGr Unltl 

*Trrlllng unita 

Strndr.dr DeveloPmont 

aHew Y.hiclc 

-Corhpontnt Ind rYitril l.Yrl 

rln-rrrvlcr Ychlcle 

-Lofdlng rnd opfrillonrl utl 

-ContlCuratlon cholctr 

OrlYrr Eludlri 

t Srmple drlvcrt' rbllltlar lo 

trntE rollovor onial 

f O.t.rGlne roll 16lilsd Yihlclt 

barad cuts 

aArrerr illact ol vrrylng 

vahlcl6 cues on drlYiri' 

conlrol rttponr6 Piltafnl 

accidents. In 1984. two-vehicle collisions accounted 

for 37.7 percent (16,668) of all highway related 

fatalities. Collisions between medium,/heavy trucks 

and other vehicles resulted in 2l percent (3,423) of all 

the fatalities sustained by occupants of other smaller 

vehicles involved in two-vehicle collisions. The majority 

of these victims (71.9 percent, 2,461) were passenger 

car occupants. In all, these 3,423 fatalities represented 

7.7 percent of all the highway related fatalities 

occurring in 1984. 

Research Plan For Reducing Heavy Truck 

Aggressivity-Frontal Impact 

Attenuation/Override Prevention 

The extent to which the effects of collisions between 

medium./heavy trucks and other smaller vehicles can 

be ameliorated is not clear. Readily achievable solutions 

are not apparent, however, given the appreciable 

number of occupants of other smaller vehicles (3,423) 

who are killed in collisions of this type, it appears 

worthwhile to study the possibility that even small 

incremental improvements can be achieved. 

Sixty-eight percent of the fatal carlcombinationunit 

truck collisions involve the fronts of trucks. 

Logically, then, work would begin on this portion of 

the vehicle. The program to explore the feasibility of 

practically modifying truck front end designs is shown 

in Figure 5. 

lntormillon Dlriemlnlllon 

aD.lYfr tirlnlng rlds 

Vrhlclr gtudlrr 

alnYanlory In-rrfvlce fl6ol fot 

grrvtlanca of unrlible 

'Oullllrt, 

tDrtrrrnlnr ptevalonct of 

mrniuvart whlch crrito n€er 

llmlt D.rlofsinct d6hrnda 

aDatermln6 tlre p6rtormencE 

EonlrlDutory ettdclr on 

rlrbl lllY 

-Combtn6d longlludlnrl ind 

lrlfrtl trecilon crprblllllss 

-Elfacti ol werr on l.Nclion 

.Arr6it trtd6-offt tltoclsl€d 

wlth vihlclGl tpcc'd lot 

oplln'rurh rtablllty p6rf orlhrnco 

a Prrf ormanca mariuTrminl/rrtlng/ 

.apoftlng guld.llner 

rSt.bltlty rrftlflclllofis of rlr. rnd 

rrlghtr rrqulrcmrnla Ghrng€3

As s i C e r$_!_!y.qs.tlc_all o!]:[]n.6_!a_t ! q_An C[_slf 

' Establish Upper Bound ol Car Crashworrhiness Capability 

. Establish Flanga of lmpact Velocities and Closing Anglos in 

Car/Ttuck Collisions 

r Estdblish HNng6 ol Kinetic Energy That Truck would Hav6 

to Absorb 

Oec i-s-ion 

. WhEi Kindtic Energy Absorbtion Lsv6ls Can 8e Hrndled? 

. Countcrbalancs Operational Practicality Concarns Against 

Perlormance Goals 

r EstBblish Perlormance ObjectivsE 

Dec-tq'.-o-!-r 

. ArB Alte.native Oesigns Practical? 

. Do They lmorove Sdlety Pcrldtmanco? 

. Conduct In.Sgrvice EvaluEtions 

- AFs6ss Duralrility, Pcrlormance, Etc. 

Figure 5. Truck frontal attenuation/aggressivity reduc' 

tion research Program 

Truek Occupant Crash Protection 

Each year about 1,000 occupants of medium and 

heavy trucks are killed and 400,000 injured in crashes. 

Most are drivers of combination-unit trucks. Highway 

crashes are the principal occupational hazard faced by 

truck drivers. 

Accident data analyses have indicated that not all 

heavy truck occupant fatalities occur in catastrophic 

accidents. Rollovers, cjections, entrirpmellt in crushed 

cabs, contact with itrterior surfaces, and fires are the 

primary mechanisnts responsible for the majority of 

truck occupant fatalities. Most of these fatalities 

occur in sirrgle-vehicle accidents which involve running 

off the road and hitting fixed objects, simple otr-road 

overturning, jackknifing (with or without overturning), 

and collisions with low roadside structure$ $uch 

as guard-rails, sign posts, and emhankments. 

In many crashes which are now fatal to truck 

occupants, use of a safety belt more than likely would 

have been all that wa$ necessary to avoid the fatal 

outcome of the accident. No other L'ountermeasurc is 

likely to reduce thc number of truck occupant deaths 

and serious injuries a$ greatly as would increased 

re$traint system use. Beyond this, further improvements 

could be realized through tntck design$ that 

provided: 

Seats and restraint system$ that keep the 

driver firmly in place and prevent him from 

being thrown around inside the cab or being 

ejected, 

SECTION 4. TECHNICAI, SESSIONS 

I A reasonable amount of protection from 

post-crash fire, 

t Cab interiors free of sharp, hard objects that 

can cause injury during impact-especially 

the steering wheel rim and hub, 

t Strong, rigid cab designs that provide occupant 

survival space in a crash (within reasollable 

expectations), and after a crash, means 

of escape. 

Before practical solutions can be sought in any of 

these areas, both the types of crash cnvironments for 

which protection would be sought, and atr upper 

range of severity lbr which practical solutions are 

deemed feasible, would have to be established. Using 

this approach, multiple-event collisions/impacts represent 

the most reasonable group of accidents to address, 

while those involving high-speed impacts into 

rigid non-yielding objects (such as bridge piers) clearly 

are not. lnitially, 9 g crashes would be assumed to be 

the maximum level for which improvements could 

practically be sought[5]. Further attempts to refine 

this estimate of a reasonable upper bound of deceleration 

woulcl have to await the completion of detailed 

recon$truction and analysis of accidents that have 

been investigated in-depth. 

Research would proceed along both analytical and 

experinrental lines. The analytical work would consist 

of in-depth accident investigations and mathematical 

modeling. 

The accident investigations would be targeted to 

better definc typical crash-induced forces, direction of 

force application, luel and ignition sources in accidents 

involving fires, occupant trajectories, and causes 

and amount of occupant injury trauma sustaincd. 

The modeling activitics would focus on delining 

vehicle dynamics both before and during the crash 

event, describing the response of the vehicle structure 

to crash-induced loads, and predicting occupant 

trauma outcomes (given assumed crash forces). 

Existing approaches and/or standards woulcl be 

used wherever possihlc. Component-level testing (as 

opposed to full systems-level tests) would be the 

preferred method for experimental work. Existing 

standards (for example, those now used for rollover 

protection systems (ROPS) for construction equipment 

and the Swedish and ECE cab structural irrtegrity 

tests) would serve a$ initial reference points l'or 

experimental testing. 

The emphasis on all this work would be to achieve 

practical increnrental improvements rather than strive 

for total, yet unattainable, solutions. 

Research Program To Improve Heavy Truck 

Occupant Restraint Systems 

Observational surveys of seat belt use among 

combination-unit truck drivers indicate that few (6.25 

651

percent) use them[6]. Surveys of truck drivers' opinions 

relative to safety belts indicate that many erroneously 

believe that they are not cffective in reducing 

injuries in crashes. ln addition, many drivers feel 

safcty belt designs are deficient. Drivers said safety 

belts were often dirty (because of lack of rctractors) 

and uncomfortable because they did not ',give" in the 

truck's relatively rough riding environment. 

Efforts to improve this situation have taken several 

tacks. First, an industry/government ad hoc group 

developed a packaged safety bclt use promotional and 

information kit aimed exclusively at motor carricrs 

and truck drivcrs. The package has been promoted 

through the American Trucking Associations' 50 state 

trucking organizations. 

Secondly, NHTSA issued an Notice of Proposed 

Rulemaking (NPRM) to amend FMVSS 208 to require 

emergency locking retractors (ELR'S) and pushbutton 

releases on truck and bus salety bclt asscmblies. 

The purpose of the proposed amendment was to 

promote the use of saltty belts by cnsuring that they 

are comfbrtable and easy to use and that fhey remain 

clean. 

Truck safety belts without any retractors often 

become entangled in the suspension mechanism of 

truck seats where they become dirty and difficult to 

extract and use. It was rcasorred that by requiring 

ELR's, safety belts would be kept clcan while ensuring 

that they not cinch up around drivers' waists in 

the comparatively rough riding environment of a 

truck. 

F'inally, eft'orts have proceeded to f'urther upgrade 

the performance and comfort of truck safety belt 

designs. There is growing interest among some truck 

operators in having workable, comfortable 3-point 

re$traint systems in heavy trucks. Research is needed, 

however, to determinc how best to modify and adapt 

3-point systems to the ride environment typical in 

heavy trucks since pa.ssenger car ,$y$tems cannot be 

uscd in unnrodified form. Manufacturers also report 

that their efforts to design 3-point systems for hcavy 

trucks are often constrained by restraint system 

strength requirements they feel are unnecessarily high. 

Accordingly, the primary thrust of engineering research 

to improve restraint sy$tem$ should be directed 

towards ascertaining strength requirements for truck 

safety belts that are appropriate for truck occupant 

protecti()n goals. These levels should be bascd on the 

objcctives of preventing occupaflts from being thrown 

about inside the cab while at rhe same time preventing 

them from being ejected. The objective would be to 

achieve a balance which optimizes occupant protection 

performance while also enabling maximum clesign 

flexibility. 

The program is shown diagrammatically in Figure 

6. 

652 


Orllnr T.!Ct lld! Con.trrthtl 

a Ertrbltrh cofrtott thrtthold 

Ylt-.-vlr b.I ctnchtnO 

Erl.bllrh T.ntrtlv. l{iw Etr.ngth 

FcllorhFnce Lsvala 

Trrt ol Btvtlrtt Byrtror 

lal{hr Prrdlcllon Analyrar 

aPlr occup.nl tlnulrlot 

I B'ad lalt!-Fronlal crarh rrr!irfitni 

tClb roll liilr-Foltoy!r crrrh rtraaamanl 

In B.rvlc. Flatd t.rtr 

aVcrlly comlort, !ors-ol-uto 

frodolr lq Sug{ lr0trctlon 

tltprdcd rt tttrrnrttvt b.lt 

llrlnglh IrYalr 

Figure 6. Truck occupent restraint systems research 

program 

Research Program To Prevent Post-Crash 

Fires 

Post-crash fires are involved in nearly l6 percent of 

all fatalities to the occupant$ of medium and heavy 

trucks. The comparable figure for passenger cars is 4 

percent. Studies that have been conducted on this 

issue indicate that between l0 and 50 percent of these 

fires result from fuel loss lrorn the truck's I'uel 

containment and delivery system [7]. 

Fire related accidents are obviously very severe. In 

many cases, it is impossible to determine whether the 

crash events precipitating the fuel release wcre so 

severe-in and of themsclves-as to preclude vehiclebased 

upgrades. Nevertheless, reasonable incremental 

improvements may be possible. 

Accordingly, a research program seeking ways of 

preventing post-crash truck fires would be directed 

towards enhancing the ability of truck fuel systems to 

withstand crashes and remain intact. It would include 

investigations of both the feasibility and desirability 

of using: "kill switches" in the truck's electrical 

systcm to eliminate this a sourcc of ignition; bladders 

in fuel tanks to contain fuel in damaged tanks; 

non-toxic, flame-retardant materials in truck cabs; 

on-board firc suppression systems, and; fuel delivery 

and return lines rerouted away from engine and,/or 

exhaust heat source$. The program is shown diagrammatically 

in Figure 7.

Fud|-PallY try $yrtam 

T h.rftr I Antlyrlg 

I Harl rour( 

tuel lampc 

t Trrdi-olf 

*lutl rlr 

al. k 

-dfrlgn rl 

at rrlrlng 

tturoa 

Intlyil6 

ng YE lire 

arnallYtl 


Aceld.nt Raconrlrucllon/ Anttyrtr 

I Cralh nrchanlrmr cacalng ayrlam br.achlng 

I Crrth lorcar/allracllon of rppilcrrton 

a lgnlllgn tgurc.i 

Figure 7. Truck fire prevention re$earch program 

lmplovf( lD..lgn 

Alltrnrliyaa /F.rrlbllltY 

a H.loctrlon 

Ana I lalr 

t Shi.ldlng/ rlranglhihin0 

I Fuil llni I 

fOutlnq 

9rclalon 

llcfnttlY6 

a Incrrrhitrlal datlgtr 

chrngrr llh.ly to yl.ld 

rrlrty lmpfdvrrrnl? 

Yrr 

OaYalOp/Frbrlcric Protoltpaa 

tFlrld lrrl fori 

- l{|nuf rcl urrbilf ly 

-P;tcl lc r llt y 

-Srfofy paito.manc. 

Research Progrnm To Improve Steering 

Assemblies And Other Cab Interior 

Components 

Serious injuries sustained by truck occupants who 

remain inside the vehicle in accidents (where intrusion 

of the occupant compartment is not sufficient to 

cause entrapment) result from contact with interior 

components. Where specific injury information is 

available, the steering wheel has consistently been 

identified as the primary sourse of serious injury to 

heavy truck occupants, followed by sun visors/roof 

DrYrlop Srfcly 

Ptrlorntnct Mciaurtmtnl 

Pr ocrdural 

9oaDponant^Lavf l Trrl! 

t Erlrbllrh mrrlmum 

Parlorminca crprbilily 

vlr: 

llo 

-Drop taili 

-Abrr1ls6 1"a1a 

-Punctura tttll 

Errk glher 

oOunltrmltlurtr 

top moldings, roof and side rails, and instrument 

panels. 

Analyses of occupant trajectories in accidents reveal 

that the majority of crash-involved drivers move in 

more than one direction durirrg the crash sequence. 

This is possible because many truck crash events occur 

over long time periods (seconds) compared to those of 

cars (milliscconds). Given this much movemcnt by 

occupants, it is not suprising that besides impacts into 

steering wheels, contacts with the windshield, instrument 

panel, and surfaces of doors and door headers 

653

have also been identified as common sources of 

injury. 

The goal of this research program, therefore, would 

be to develop means of reducing occupant injuries 

that result when truck drivers, both restrained and 

unrestrained, impact cab interior surfaces and components, 

especially truck steeritrg wheel,/column assem- 

The program is shown in Figure 8. 

Research Program To Enhance The 

Structural Integrity Of Truck Cabs 

A crash performance objective that heavy trucks 

share with pa$senger cars is maintenance of sufficient 

space within the occupant compartment to "ride-out" 

the crash event. This capability is of equal importance 

with that of restraining/containing occupants within 


Figure 8. Truck cab lnterior components research program 

654 

Ocouprnt Trluma 

Pr.dlcllon Hodrla 

I Utr truch rl.crlfig 

shatl grsmrtrlrt 

t urr dacalaartlonr 

< 0g'r 

In-DrDlh Ascldrnt R.oonatructlon I Anrlyrlr 

t Typ. of Inlury rurtallfd 

t F.glon ot bodt 

t Prrtr ol botlt ilorl ollan Inlurrd Dy rtfttlne 

r..tmbllf. 

ll Int.rlor componfnlf 

Orrlgn Dfvllopfilnl 

"8ollrr" 

Slrd Tr.tr U.lng ATD| 

Syrt.mrllcrlly Yary 

dfcalttallofi rtttt, 

mil.urlng lmptctlflg 

fcrcar I altrrlng rh.rl 

dafornatlona 

Dttrrilln. Irtunt gau.fd 

by tltrrlng rh.rlr rnd 

olhar conDonanlr 

O.c I rlon 

t Oo rccldanl rrconrtructlon 

tnrlyrlr. llrd lrrtr rnd 

lrb ecftpr.rrlon lttlt 

lndlcrlr porrlbllltt to 

lmprov. d.rlgnt, 

pritoamtnct? 

crD Inl.rlor 

the vehicle since it logically follows that if ejection is 

prevented it is of no avail if the occupant is subsequently 

killed or injured due to the cab crushing. 

Therefore, the objective of research on heavy truck 

cab structures would be to determine if presently 

available cabs could be strengthened practically to 

enharrse their ability to protect occupants in noncatastrophic, 

yet potentially lethal crash environments. 

Excluding those ejected, 20 percent of all fatally 

injured combination-unit truck occupant$ are extricated 

from their cabs-a surrogate indication that cab 

structural integrity is, at least partially, involved. 

European (Swedish and ECE) standards exist 

which, if applied to U.S. trucks, would result in 

incremental strengthening of most U.S. designed truck 

cabs. Research is needed, however, to determine if 

practical modifications can bc made to U.S. cabs that 

F.tlnr Trrrlng llrthodr 

Controllfd SlrrlG 

Lrborrlort Trrt. Ot 

Elilrlng whmlr 

a Har.ur. tcrca, 

datohrllon 

chrrrctrrlrllc., ol 

tlatrlfig :h..1 ?lm. 

and tun vlrort, dt.h 

Dtnalt. alc.

would materially improve the likelihood of saving an 

appreciable number ol' victims in typical U.S. truck 

crashes, especially rollovers. Limitations and tradeoffs 

are inherent in such an endeavor and must be 

recognized and acknowledged at the outset. The 

program is shown in Figure 9. 

Summary 

In the U.S., medium and heavy trucks are annually 

involved in crashes which result in over 5500 people 

Ftan, Ettht^t Ahatrt.a 

In-Drpth ACCldanl nacsnriructlon and An.lrtta 

a lmprcr lp.cr htg6llud.r 

I Olf.clronr In Fhlch torca la aCpttad 

a O.h.0r r.v.ilry 

r Tr.vh. n.chrnlrmt 

a;orc./d.lbctlcn ch}rFh,laltcr lol 

vrrlaur dlrrctlOfra ql lorca 

Orvrlopnad ol rn lmprova{ 

C.b Srrccturr 

lC9ghlrrnl ol ralchl/ilrrnslt 

a cognlrarr ol cs.r lFprrcarrorl 

. CoChtrrnt ot nrrd tA daatCn In 

arrrh Irol.clro^ lr.lv.ta lor olh.r 

fanl{lt. hll by t/!€lr 

Figure 9. Truck cab structural integrity research pro- 

9ram 


European Review of Heavy Goods VehicleSafety 

L Neilson 

on behalf of the Ad-hoc 

D.v.lopn.nl ol Slnplltl.d Conponrnl 

Lrvrt Labqrrtory Trrta lo.: 

a $tructurrr 

twtndrhtrtd 

r ooor./rlllarlalchl 

Group of the European Experimental 

Vehicles Committee on Side Impact 

Dummies, 

United Kingdom 

Preface 

The Comrnittee of EEVC (European Experimental 

Vehicles Committee) decided at their 1984 policy 

; 

being killed and 170,000 injured. There are many 

complex and interrelated rcasons why these crashes 

occur, among them being lactors which relate to the 

way the vehicles are designed, maintained, and perform. 

This paper identifies the key vehicle-related 

medium and heavy truck safety issues and briefly 

summarizes programs of research to achieve improve- 

ments. It is hoped that these research agendas will 

serve as blueprints for both industry ald government 

efforts to achieve 

those improvements. 

References 

l. 

"Heavy 

Truck Safety Study", Congressional Report 

prepared in response to Section 216 of the 

Motor Carrier Safety Act of 1984, P.L. 98-554, 

yti::htlt;ck 

occupant protection", congressional 

Report prepared in response to Section 217 

of the Motor Carrier Safety Act of 1984, P.L. 

98-554, January, 1987 

U.S. Department of Transportation, Federal 

Highway Administration Annual Roadside Vehicle 

Inspection Reports 

Radlinski, R., "Heavy Truck Braking Performance*The 

State-of-the-Art in the U.S.", Society 

of Automotive Engineers (SAE) Paper No. 

870492, t987 

Clarke, R.M. and Mergel, J., "Heavy Truck 

Occupant Protection-A Plan for Investigating 

Ways to Improve It", SAE Paper 821270, 1982 

Allison, P. and Tarkir, R., "Heavy Truck Occupant 

Restraint Use", Final Report, U.S. DOT 

Contract No. DTNH22-80-C-07451, Septernber 

1982. 

O'Day, J., Ruthazer, R., Gonzales, T., "An 2. 

3. 

4. 

5. 

6. 

7. 

In- 

Depth Study of Fire Accident Data" University of 

Michigan Transportation Research Institute Report 

No's. UMTRI 84-40 (1984), and UMTRI 

85-17-r (r985) 

meeting that an informal Working Croup on Heavy 

Goods Vehicles be set up to consider the road 

accident situation in Europe for these vehicles. It 

should also report upon the progress being made to 

improve the accident avoidance capability of such 

vehicles, the protection that would be provided for the 

crew ancl the protection that might be possible for 

other road users involved in accidents with Heavy 

Goods Vehicles. This informal Working Group did 

not start work until 1986. It was requested by the 

555

main EEVC committee in 1986 that it report in time 

to present papers to the OECD Symposium on this 

subject in April 1987 in Montreal, Canada and to the 

llth ESV Conference in May 1987 in Washington, 

D.C., U.S.A. The former objective has not been met, 

but the pre$ent report in preliminary form and without 

the final approval of the main EEVC cornmittee is 

now presented. 

The informal group consisted of: 

Mr Pullwitt, lBASI, FR Germany 

Mr Cesari, IINRETS, France 

Mr Fline, IINRETS, France 

Prof. Strandberg, lVTl, Sweden 

Mr Tromp, ISWOV, Netherlands 

Mr Riley, ITHRL, UK 

Mr Neilson, ITHRL, UK (Chairman) 


In several parts of the world there has been 

increasing concern about the contribution that Heavy 

Goods Vehicles, and similar vehicles designed for 

special purposes, are making to the overall road 

accident situation. The prcsent study shows that 

generally speaking the situation in Europe is somewhat 

improving for a complicated set of interrelated 

reasons. This report is intended to review the situation 

and to discuss engineering means by which the design 

of these vehicles can be further improved. This is a 

continuing and important objective because in some 

countries l.here has been a widespread fear of these 

large road vehicles and the accidents and injuries that 

they can cause. This has been hindering the development 

of road transport and has been leading to 

restrictions on the operation of such vehicles in some 

places. 

Any discussion of heavy commercial vehicles should 

start with some definitions about which vehicles are 

included and how the sizes are divided up. The 

present study is concerned with Heavy Coods Vehicles 

and this excludes light goods vehicles, bus and 

coaches and misccllaneous large vehicles which do not 

carry goods. The point of division betweerr light and 

heavy is not the same in different countries in Europe 

ancl varies between about 3.5 and 7.5 tonnes Gross 

Vehicle Weight. Taking the lower limit, HGVs are 

about 390 of the vehicle population, but they may 

cover 890 of the distance. Othcr large and specialist 

vehicles which do not carry goods are an addition to 

the number of vehicles, but they appear to have a 

much lower involvement rate in accidents presumably 

because these vehicles cover much shorter distances 

each year. The following section gives an indication of 

the part that Heavy Goods Vetricles play in the overall 

road accident pattern in one country. It is mostly 

based on the accident statistics for Great Britain. 

656 


The Accident Situation in General 

Terms 

The Heavy Goods Vehicle fleet in Europe may be 

made up of about 600/o two axle rigid trucks, l09o 

rigid trucks with more than two axles and 3090 

articulated or full trailer vehicles. The traffic and 

distance travelled by these vehicles has been increasing 

slowly during the past ten years with the rather 

variable economic situation and the increasc may be 

only about l5Vo over that period. However in some 

countries their involvement rate in accidents per 

distance covered has dropped by almost a third and 

the fatal injury rate for their drivers by a larger 

amount (possibly this has almost halved). These 

reductions are additional to the rather similar large 

reductions which occurred in the previous ten year 

pcriod up to about 1975. 

Almost all casualties in accidents involving HGVs 

are to other road users rather than to the drivers and 

passengers in them. These casualties are mostly divided 

between car occupants, riders of two wheelers 

and pedestrians with the HCV occupants accounting 

for only about a tenth of the total. The implications 

of this situation are discussed later on in some detail. 

For example, it has been noted that a proportionally 

large number of latal pedestrian casualties occur in 

accidents involving articulated vehicles. Presumably it 

is the sides of these vchicles which cause additional 

fatal injuries. 

Most HGV accidents leading to fatal injuries to 

their occupants occur away from built-up areas but 

for those injuring other road users perhaps a third 

occur in built-up areas. lt is noteworthy that most 

fatal accident$ outside built-up areas occur on the 

main roads, whereas in built-up areas there are a 

surprisingly large number on minor roads, It is the 

larger vehicles within the HGV category which have 

many of their accidents outside built-up areas and the 

two axled vehicles which have relativcly more accidents 

in town$. Another way of corrsidering risks to 

HGVs on the different roads is to compare their 

accidcnt rates per distance travellcd. By this measure, 

if the rate fbr roads outside towns is taken a$ 

standard, then the ratc on Motorways is about a half 

of that, but the rate in towns is only slightly higher 

thart outside them. 

About half of the accident$ to HGVs occur at 

junctions, but with rather fewer at junctions tor the 

largest vehicles, It is these vehicles which have rather 

more accidents away from junctions and this largely 

accounts for the diff'erence. Skidding of HCVs tends 

to be reported when control is lost of vehicles and 

they depart from their intended path. lt is also 

reported when tyre marks are deposited on thc road 

surface. A half of reported skidding accidents occur 

on wet roads, but there are almost a$ many in dry

conditions. In most countries the totals in snowy and 

icy conditions are relatively small. 

Road accidents occur in many different circumstances. 

A comparison with the manoeuvres of cars 

before accidents shows that HGVs have relatively few 

accidents, presumably because their drivers are professional. 

They have particularly few when waiting but 

held up, when turning to the offside and when just 

going ahead. However for accidents when going ahead 

at bends when overtaking, when changing lanes, when 

reversing and when parked, their accident rates approach 

those for cars. These findings rather suggest 

that HGV drivers have real problems in seeing obstacles 

and other vehicles but not when turning across 

traffic at junctions because they have a good view to 

the ofl'side. 

The main sections of this report now lollow. There 

is a detailed comparison between the accident statistics 

for several European countrie$. Then there is consideration 

of the rernedial measure$ lor HGVs which are 

becoming available ctr which appear to be required. 

Firstly there are accident avoidance features such as 

improved braking. Then there is a section on protective 

measures. This is divided between protection for 

the occupants of HGVs and protection for all other 

road users likely to be involved in accidents with 

HCVs. 

Accident Situation 

For a comprehensive assessment of the European 

accident situation involving heavy goods velticles 

(HGV), it is nccessary to come to commou definitions 

of HGV weights, sizes and other aspects. 

It is reasonable to use already existing common 

definitions as far as available, e.g. withitr the Directivcs 

of EEC, where r-rrtil'orm definitions for HCV 

sizes, weights and axle loads will become effective 

gradually by l99l (85/3/EEC). 

This report discusses only vehicles for the transport 

of goods with a gross weight of more than 3.5 tonne$. 

As far as possiblc the corre$ponding figures for 

different countries are split up in this way. Figures for 

buses and in special cases for cars too, should 

underline the comparisons for specific situations for 

HGVs. 

In Appendix I a detailed comparison is given 

between the national regulations of Great Britain and 

Germany. The use of the international provided rules 

shows that in geueral r-rc'.rrly the same differences exist 

between the classes of speed limits and driving licences. 

In Creat Britain there is a more varied 

classification for HGV licences and there are higher 

maximum speed limits for HGVs. ln ordcr to compare 

the accident situation in the European states 

contributing to this report, it would be necessary to 

consider details of the relevant figures and of accident 


rates and special accident risks. In Table I a comparison 

is made over eight years of European figures for 

registration and rnileage of HCVs (gross weight over 

3.5 tonnes) to give some information about the 

increasing presence of HGVs on European roads. 

The trend of slight increasing figures of usage is 

combined in ncarly all countries with dccreasing 

accident figures. This reflects in principle the eft'ectiveness 

of rules and regulations in the area of road 

and road equipment, training and monitoring of the 

drivers, the regulations for vehicle design and its 

supervision. 

ln Appendix 2 a review is made for Great Britain 

and Germany about the development of national 

roads with reference of increasing length and width of 

different road categories. These figures show a further 

possible reason for the decreasing number of accidents, 

namely the increasing length of Highways and 

Motorways and the widening of other roads. 

There is also the possibility of a great influence of 

technical improvements made on HGVs and of obligatory 

technical inspections upon thc accident figures. 

Technical improvements for safety are di$cussed later 

in this report. In Reference 3 and Appcndix 3 some 

European figures are available for an estimation of 

the technical inspections made on HGVs which suggest 

that in Germany the number of dcfects found on 

trucks has declincd, although this has not happened 

for cars. 

All those important factors of road traffic in the 

countries considered cannot give a complete reason 

for the common trend of decreasing figures of fatalities 

and injurcd persons, as showtr in Table 2. 

Other road users are endangered to varying extents 

by trucks in the event of accidents. This circumstance 

can be studied in Table 3, where the fatalities are 

shown by category of road users in HGV accidents in 

a European comparison. The largest trumbers oI 

persons killed in the countrics listed are in accidents 

of HGVs versus cars; this is certainly caused by the 

great number of cars in traffic. But there are obvious 

diflerences between France and Sweden on one side 

and Great Britain and Cermany on the other. In 

France and Swedcn with a greater amount of traffic 

outside built-up areas it seems to be that accidents of 

HGVs versu$ cars produce injuries of greater sevcrity; 

the lack of figures for comparisotr does not allow a 

more concrete statement. The same circumstances 

probably give a reason for the low percentage of 

injured HGV occupants in Germany in comparison 

with the others. 

For Germany the low figures of injured or killed 

HGV occupants might be also founded on the high 

mileage on highways and the high safety level of such 

roads. 

657

Table 1. HGV populatlon and dlstances travelled 

Coun try Gross weight = 3,5 t 

Frrnce Iruck s 

Roed User I 

(Rieid) 

Tractor s 

(Articulated 

Vehicles 

) 

Gemany Truck s 

Great 

Eritain 

SYeden 

(R19id) 

T rdc to rs 

(Artlculated 

Vehicles) 

Iruck g 

(Rield) 

T ra c tors 

(Articuldted 

Vehicles) 

,or, t 

Indication 

Number of vehicles 

Di s idnce travel led 

(i{ill. km) 


Distance treYelled 

(Mill. kn) 

l{umber of vehicles 

0lstrnce trtvel led 

(t'lill. kn) 


0istdnce travel led 

(Htll. kn) 

llunber of Yehlcles 

Distence travel led 

(tllll. lnr) 

Hunber of vehlcles 

0istance trrvel led 

(t'ti 

l l, km) 

l{umber of vehicles 

0istdnce trrvel led 

(llill. km) 

EXPERIMENTAL SAFETY VEH ICLES 

I - . . 

with or without trailers (semi-trailers), 

estimated on lst January, 1985. 

I 976 l9 78 1980 I 98? I 981 

501 746 

ll 189 

130 105 

3 657 

433 3oo 

ll 396 

lo? Bbo 

5 477 

6l? 221 

13 825 

I47 629 

3 689 

345oIq 

* Motor HGVs (not trailers) with maximum permissible 

gross weight above 35OO kg, registered on Ist January. 

The proportion of killed pedestrians is nearly the 

same in Great Britain, Sweden and Germany but in 

France it is clearly lower. Compared with Germany 

and Sweden and to a le$ser extent in Great Britain the 

percentage killed of the unprotected road users is very 

low in France-but here we have the highest absolute 

number l'or motorized two-wheeler riders killed. 

ln the Nordic countries the accident risk increases 

more for HGVs than for cars when the road surface 

becomes more slippery. Snow or ice is a major 

environmental factor also in absolute figures, since it 

was present in about every second HGV accident 

during a whole year period according to data from the 

Swedish National Road Administration. It can be 

assumed for HGVs that the low coefficient of friction 

and the inferior yaw stability in association with poor 

brake force distribution causes an overrepresentation 

in accidents on slippery roads. 

The problem of inferior yaw stability is also a 

problem in France, where a lot of traffic is on roads 

with insufficient width for driving manoeuvres. In the 

United Kingdom there is more concern about the 

braking power available on wet roads at high speeds 

but irccidents on urban roads are also important. 

658 

384 8oo 

ll 164 

lo4 goo 

5 953 

207 obI 

t2l 893 r33 4?6 ' 

559 169 

ll 765 

l7l 143 

3 9t4 

374 9oo 

lo 455 

99 7oo 

5 4t4 

t9 o79 8? 016 83 3?t 

? 8o5 

640 ??l 

l3 509 

l8B 400 

4 65/ 

l4 7 4oo 

lo 425 

88 loo 

5 569 

6?t 531 

l3 ll4 

209 179 

4 89t 

3{6 5oo 

lo 38? 

90 7oo 

6 o2? 

82 464 85 609 

In Germany serious problems are given by the 

possibility of high speed driving on highways (Autobahn). 

Investigations of BASt in 1985 give some 

figures for the speed of HGVs. The background for 

these measurements was equally in relation to roads 

and road equipment, traffic volume and so on. The 

Investigationl resulted in tbllowing figures: l69o of all 

HGVs drove within the allowable speed of 80 km/h, 

about 6090 had a speed between 80 and 90 km,/h, 

2090 were measured with a speed between 90 and 100 

km,/h and TVo ran over 100 km/h. Almost every 

fourth HGV was faster than 90 km/h. 

In connection with inclemency of the weather this 

high speed level can cause several accidents with a 

large number of vehicles involved, as occurred in 

Germany in 1985. 

Safety Measures 

As pointed out above, official statistics from many 

countries indicate that Heavy Goods Vehicles are 

overrepresented in fatal accidents. In fact, the absolute 

numbers of fatalities involving HCVs are alarming 

enough to motivatc special research on this 

category of vehicle and road user.

Table 2. Comparlson of eccident flEures In Europe 


3oun t ry Iddicrtlon t976 t978 | 980 | 98? ' 984 

Fra nce 

Germilny 

Gred t (l 

Eritdln't 

St*den 

Accidents Hlth 

personrl lnJurles 

HGVs lnvolved 

Fatrl i ties 

HGVs involved 

lnJured occupnntg 

l6Ys involved 

Accidents Hlth 

personrl inJurles 

HGvrll lnvol "ed?l 

ACcldents Hr th 

fatalltles 

,tGYsl) invol vedl I 

ACCTOenE lnv0,Ytn9 

rerlous rnd slight 

injury 

i.tcvrll involveo?l 

Acc i denti Yi th 

persgprl inJuries 

HGVrU inYolvedcl 

Accidents rt th 

fdtdl ltler 

HGYs ll i nvol vedd 

Accident involving 

serlor.lr rnd rl ight 

ln lury 

xGls rl Invot ved?l 

Nurber df dccidentl 

Yl th I nJuri er 

trr:v.lli.u^t'oa4l 

luntrcr of rccldentg 

ri th frtrl I tlei 

tGvsjl lnvolved 

359 694 

?7 qlA 

t3550 

I 7tE 

146 t44 

?5 223 

380 35? 

lf All HGYs rithout reight llnlt (ln GB, goodr vehlcler grertcr than 1,5 tons 

unladen weiqht) 

21 0nly dccldents ylth odr rnd t|f Otrtiet Inyolved 

3) HGY rlth mrxirun pernlrsr$lr Gross Vehlclr tlrlght rbovc 3,500 tg 

4l lluntcr of Involved (not only prlnrrlly) vehlclrg 

5| Engl5nd, Scotlrnd and Yrler snly (dorr not lncludG llothern Irrlrnd) 

6) Thesr trt nunters of dccldrntt lnvolvlng HGYr 

379 ?35 

?R t ol 26 ?97 

t3 368 tl 9tt 

355 984 

26 455 

258639 ?64 769 

rl 

t1 F00', r3 858 

6 006 6 30f 

llo6l 752 

?526t3 

t58 465 

tl 

ll 

030 

041 

I ?88 

I 035 

Though a general decline may be found in the total 

number of traffic accident fatalities. the relative 

aggressiveness of HGVs (when compared to cars or 

other motor vehicles) seems to be comf)aratively 

constant. In fact, data from a U.S. review 1985, "Big 

Trucks". from the Insurance Institute for Highway 

safery, llHS (waterBate 600, washington D"c.) exhibit 

a slight increase of HGV aggressiveness in the 

last decade: "ln 1977, a car occupant was 26 times 

more likely than a truck occupant to be killed when 

the two vehicles crashed. Now the ratio is 35 times 

more likely." 

During the last decades, some international seminars 

and multilateral reviews have been devoted particularly 

to HCV safety, for instance by HSRI (1975, 

later University of Michigan Transportation Research 

Institute UMTRI, Ann A,rbor), OECD (1977), and 

l8t 

r3 r06 

l5 0?8 

I t{5 

9t I 

l6? 

I 454 

361 3?4 

?4 ntr 

?52 J00 

lt 4l/ 

5 550 

6?l 

?46 710 

t0 791 

r5 ?3t 

I t3e 

755 

tzl 

?.23l5? 20? 0t5 

il 509 

r2 102 il 685 

I t73 

3r?8?Z 284 905 

358 593 

22 764 

t0 581 

L z?3 

3{E a 

?l 011 

?55 980 

r0 $89 

5 441 

620 

?50 533 

I 969 

t5 288 

I 0t9 

681 

l15 

t5 27? 

159 485 

]7 Anl 

I t04 

I I ll 

J5U t6t 

?l 7?8 

?53r83 

t0 t7l 

5 t38 

591 

?48 015 

I 580 

t6 531 

I n17 

OECD (1983). More recent studies on HGV safety 

have been (NHTSA, 1986) and will be published by 

the National Highway Traffic Safety Administration 

(NHTSA), since 1985 arranging a special session on 

HGVs in its bi-annual Technical Conferences on 

Experimental Safety Vehicles (ESV). 

The operating conditions of HCVs are quite different 

from light road vehicles. In addition, the great 

dimensions and great (variations in) weight of HGVs 

have lead to design principles deviating substantially 

from cars. Many of these HCV peculiarities deteriorate 

their safety perfbrmance to an extent that probably 

would not be tolerated in more commonly known 

road vehicles. 

Safety-related quantities and possible technical 

countermeasures have been studied experimentally, 

and sometimes in real traffic, by a number of 

ltl 

il9 

6s9

Table 3. Caeualtles In aceldents involvlng HGVs 

Country Sever{ ty 

of 

I nj ury 

Frenct fatrl 

Ser i OUS 

5light 

Gennrny fatal 

ser10u5 

slight 

Grea t 

Eritdin 

fatal 

serlous 

sl ight 

s*eden?)3) fatal 

serious 

sl iqht 

pedes*rrns 

r pedrr 

cy(lrsti 

I 


l) HGV 

vtth edch gross relght. excluding accidents rith more then tHo ptrt{cg {nvolved. 

Number 

of k{lled persont ld ecciderttslhere " goods yehlcles (imesptctlve of r.eight) 

rrra J) prinar'lly lnvolyed during lg85 . 

4)ln eccldents involvlng lGVs over 7,5 tons grosr Hetght. 

, ;:"::,.r';:::''"' I rcy r 

tHo-wheele'5l occuoants occunrnts 

'rhers I rot.r (roorl 

I 

| 

I I 

No t l{0. T llo. Ho_ t llo. I l{0 t 

I44 I,l 79 4,5 176 9.9 Ir54 65.1 159 9,0 6l 3,4 tTtJ 

t76 

to4 

8s5 

104 

261 

463 

td \ 

q R 

5,0 

tq r 

8,0 

167 

803 

l5l2 

$z 

184 

366 

t1 A 

ll,2 

A A 

7,5 

4.1 

ll9 

I 058 

t724 

84 

'lql 

526 

investigators in Europe. Below, a few of them will be 

ref-erred to directly or indirectly. However, in the 

EEVC HGV Working Croup we hope that our 

unintended ignorance of other studies will stimulate 

the investigators in question to contact EEVCrepresentatives 

and to submit papers to the HCV 

session in future ESV conferences. 

Accident Avoidance 

Involved Institutions. ln the HGV session of the tenth 

ESV Conference 1985, original research results on 

HGV and bus accident avoidance properties were 

presented from European institutions such as: TUV 

Rheinland (Institute for Traffic Safety) in FRG; 

Renault Vehicules Industriels in France; Cranfield 

Impact Centre in U.K. From other publications it is 

known that important experimental research on the 

active safety of HGVs also has been conducted 

recently: in FRG at TUV Essen and at the Technical 

Universities of Aachen, Berlin, Braunschweig, and 

Hannover as well as at HUK-Verband (insurance 

company cooperation) in Munchen and at Daimler- 

Benz; in the Netherlands at the Technical University 

of Delft; in Sweden at Road and Traffic Research 

Institute (VTI, S-58101 Linkoeping). 

Analyses of HCV active $afety problems in real 

accidents have been reported by TRRL (Transport 

and Road Research Laboratory) and by MIRA (The 

Motor Industry Research Association) in U.K. as well 

as by the VTT (Technical Research Centre) in Finland 

through their Road Accident Investigation Teams. 

660 

il,5 

14.9 

10. t 

l?,2 

ll.7 

6.2 

618 

1Rfr7 

ll3?4 

1qt 

I 589 

4487 

5I,0 

53,9 

65, I 

5l,l 

50,4 

s?,9 

70 

481 

I0?9 

56 

517 

I 589 

q F 

6,7 

5.0 

I,l 

l5 ,4 

I8, 7 

4l 

?55 

655 

4l 

?9e 

l0$0 

3,4 

3,6 1,8,iiiiii 

5,0 

n q 

l?,4 

4) 

589 

4) 

ll49 

848 I 

1l 

2l I4,4 l7 Il,6 J 3.4 8l 55.s ?I? t{,4 I 0,t 146 

Accident avoidance contributions to the HGV session 

of this llrh ESV Conference have been submitted 

also by Union Technique de I'Automobile, du Motorcycle 

et du Cycle in France and by the Institute of 

Technology in Darmstadr, FRC. 

Literature references and additional hints on European 

HGV research into the active safety area may be 

found in a state-of-the-art survey by Vlk (1985, lnt. J. 

of Vehicle Design, vol. 6, no. 3, pp. 323-361) from 

the Technical University of Brno in Czechoslovakia. 

Experimental studies ancl theoretical analyse.s have 

revealed numerous accidenl avoidance problems re- 

Iated to the HCVs themselves and their load. Hence, 

it is essential to identify more precisely the vehicle 

design and operation variables decisive of HCVs' 

active safety. Such variables have been listed in 

numerous studies, and some of them will be elaborated 

on below. {See OECD papere by Strandberg, 

le87) 

Yaw Stability of Articulated Vehicles in Non-Braking 

Situations. The deteriorating influence on yaw stability 

fiom articulation was investigated in full scale 

driving experiments, computcr simulations and theoretical 

analyses irr lhc early '70s at the Sweclish Road 

and Traffic Research Inr;titute, VTl. Though articulation 

has been introduced in HGV design to improve 

the manoeuvrability and decrease the inwards offtracking 

at low speeds, it was f'ound to impair the 

handling properties and to increase the outwards 

oft-tracking at highway speeds, when the sicleslip 

angles no longer may be neglected. See fig. I. Safety

Side 

Sidesl ip 

il',='V 

fi 

0,4 

tllil ililt l|tl 

ijlj" ull L]lil -U 

70 kn,rh 

0,3 

t,2 

t,l 

f# '!ll 

1l 

-{l,l 

'7jfun/tt1 


- u,4 

Figure 1. Sldeelip angle peaks for dlfferent axles of a 

single trailer (two articulations) and a double 

trailer (three artics) HGV with the same 

overall length {24m} when making a double 

lane change manoeuver. The lateral acceleratlon 

at the mass centre of the truck or 

tractor was 1.75m/s' and independent of 

epeed. The scale for Side Force Coetticient 

(side force divided by normal force) is an 

approximate average for tyres and loads 

typical for these HGVs as measured on wet 

asphalt. Computer simulation data lrom 

Nordstrdm, Magnusson, Strandberg (1972, 

VTI report no. 9) 

h I 

G 

tr o 

E 

o 

e- 

E 

3.t 

t.0 

2.5 

t.0 

1.5 

r.0 

0.5 

0 

$lnglr 

Unlt 

measures in this context have been experimented with 

by Woodrooffe, Billing, Nisonger (1983, SAE paper 

83 r r 62). 

Though tlre lateral friction utilizatinn (or Side Force 

Coefficient, SFC: Side Force divided by normal 

force) in fig. I is at a moderate level for the triple 

artic tractor, the rear wheels are skidding at 90km/h 

with a SFC close ro the coefficient of I'riction for the 

actual road surface. During this kind of manoeuvre, 

very light braking may cause sudden and sevcre 

skidding at the rear end of the vehicle combination. 

Similar results were arrived at with experimental 

research also at the University of Michigan Transportation 

Research lnstitute (UMTRI, previously HSRI) 

and are supported by others according to the above 

mentioned survey by Vlk (1985), who reviewed 70 

articles on the handling performance oI truck-trailer 

vehicles. These experimental evidences are now supported 

by the novel results from a case-control study 

of HGV acciclents 

by Stein and Jone$ (1987) in fig. 2. 

Yaw Stability During Braking. According to schematic 

descriptions of tyre force characteristics, eq. I expresses 

the approxirnate relationship between Coefficient 

o[ Friction (CoF) and the maximum available 

(due to road friction) Br aking Force Coefficient 

(BFC = braking force divided by normal force) and 

Involvement 

of Trucks In Slngle Vehlcle 

Crscher by Truck Conflguratlon 

Tractor- 

Trrilrr 

Truck- 

Trellrr 

'frrilo ol lruck crtrh lfiwlvrfiffit pf.Etnt|ef lo cofhprflron ||mplr F.|E.nl|g.. 

lflarlarn 

Ooublr 

Flocly 

Mounlrln 

Doubfr 

Figure 2. Involvement of HGVs in single vehicle crashes by HGV configuratlon. Data lrom 222 HGVs involved in 

slngle accidents end from 666 comparleon HGVs. From "Crash Involvement ot Large Trucks by 

Configuration: A Case-Control Study" by Howard S. Stein and lan S. Jones (January 1987) with kind 

permlssion from Insurance Institule for Hlghway $afety, Watergate 600, Washlngton, DC 20037 

661

Side Force Coefficient (SFC). It is often visualized as 

the so called friction circle. 

BFC2 + SFC2 < CoFz 


(1) 

Due to the great rearward amplification of the SFC 

in a manoeuvering artic (see I1g. l), very small 

braking I'orces may then lock up the rear wheels and 

result in severe skidding, since the tyre force direction 

becomes indefinite when the unequality in eq. I 

approaches equality. The necessary side force for yaw 

stability is then no longer available. On icy and very 

slippery roads, skiddine and iackknifc accidents may 

be initiated also by traction or retarder forces. 

Skidding will occur even during straight driving, if 

the wheels at some vehicle end are braking too hard in 

relation to their load. This will bring the BFC too 

close to the CoF. Therefore, the driver has to restrict 

the brake pedal force below the level where the least 

(relatively) loaded wheels lock up. If no load' 

compensating device is installed, Table I demon$trates 

that the non-locking braking distance will increase 

more than twice when only the trailer has been 

unloaded. 

In Table 4 it is assumed that the brake force 

distribution is constant and adapted to the maximltm 

permissible gross wcight in Sweden, where no dcvices 

are required for load compensation of the braking 

torque. Initial speed: 20m/s or about 70km/h. Winter 

road with CoF:0.2. Dynamic load transfer neglected' 

These examples of unloading will lead to trailer 

wheellock if the braking force exceeds 2590 of the 

value, possible to apply at full load. 

Brnking Performance of HGV Combinations in Traf' 

fic. The handling and stability of HGVs is particularly 

poor during braking. Thereforc, a case-control study 

(similar to the above mentioned by Stein & Jones, 

1987) in the four Nordic countries was opened during 

1986 with measurements of decisivc quantities in the 

air brake systems and of the braking characteristics of 

H(iVs. In cach country, 100 HGV-combinations were 

randomly selected from the normal traffic flow on 

suitable roads. 

Data for Sweden, now being analyzed, indicate that 

quite a large proportion of the HGV-combinations in 

u$e are unable to reach the minimum deceleration 

performance required in legislation, i.e' 5mls2 with 

available air pressure when fully loaded' Wheel-lock 

and corresponding skidding tendencies (or excessive 

braking distances) seem to tre another major problem 

in Sweden, where no devices are required for load' 

compensation of the braking torque, Out of 100 

combinations investigated, 75 were completely without 

such equipment. 

In fig. 3 the deceleration values measured on these 

HGVs have been transformed to the corrcsponding 

braking distance fiom 70km/h at maximum control 

pres$ure (6bar). So have the Swedish deceleration 

requirements for HGVs and for cars. Cars are required 

to decelerate 5.8m/s2 without any wheelJock, 

and the rear wheels must not lock before the front 

ones between 5.8 and 8.0m/s2. HCVs must have 

brakes capable of decclcrating the fully ladcn vehicle 

at 5.0m,/s2 (from 60km/h). However, the Swedish 

rules corresponding to the ECE corridors (restricting 

the deceleration-pressure ratio in both directions) are 

rarely checked since the announcement of the general 

exemption from load sensing valves in the early '70s. 

More distinct and general conclusions may be 

drawn later in 1987, when data from all participating 

countries are available for analysis. Deceleration performance 

and skidding tendency will be quantified 

and compared between the four countries. Substantial 

differences in these qualities are expected, since very 

few Swedish HCVs have load compensation, while 

many Finnish ones have manually controlled pressurereduction 

valves, and while Danish and Norwegian 

legislation requires automatic load scnsing valves. 

Practical valve problems have been reported, and the 

outcome of the comparisons between countries seems 

diftlcult to predict. 

Characteristics of HGV Air Brakes Impairing $afety' 

Apparently, HGV braking properties are even more 

inferior to that of the cars than what the (quite 

moderate) demands in legislation permit. This inl'eri- 

Tabte 4. Non-skidding braking distances with ditterent loading on a typical HGV (such as the truck-trailer in fig' 

1a) without load sensing valves 

Loacl i ng 

cond i t ion 

Both loaded 6 

Both empty 5 

Tra i ler ernpty 6 

Tr.bogie unload 5 

!rpl,ghF*E*-a!"-_9he..di-Lf ere-ng-Bxle-q-*.LHnsesI 

Truck Truck Truck Trailer Trailer 

Lrs_ut Dr_i_ve ftajljlog Er-o-n! Eo.+i-e- 

I 

6 

I 

I 

I 

uP=o 

I 

I 

1 n 

2.5 

2.5 

10 

I6 

4 

4 

4 

Gross 

lve i qht 

(Esnne*s_) 

48 

17.s 

28.5 

36 

Braking 

Force Distance 

H.E,* 

1001 

fiNJ 

100 

25% 175 

zsB ?38 

25?. 300

CARS WITHOUT }.IHEEL-LOCR 

I4AXII4UH ]] H 

CARS FRONT I.IHEEL- 

LOCK ACCEPTED, 

ltHE E L* LOC K 

NO WHEEL- 

LOC K 

20 l0 

50 60 70 

HFV:S IRRESPECTIVE 

OF WHEEL-LOCK, 

|4ax 38 r-r 

SECTION 4, TECHNICAL SESSIONS 

f 

SWED I SH 

LEGISLATION 

LIMITS 

80 90 100 r10 

LOAD NEIGHT OF HFV-COHBINATION 

I GREAT El MEDtuM Etl 

Figure 3. Dlstribution of estlmated non-locking braklng sistance$ trom 70km/h ln a sample of 75. HGV trailer 

combinations (without load sensing devices). Random selection from the trafflc on suitable roads in 

Sweden 1986- Pretiminary results. Estimale based: (back row) on the smallest recorded deceleration at 

wheel-lock or (front rowj on extrapoletlon to 6bar contlol pressure from recorded decelerations in 

drivlng tests at 3ber and at 4.5bar, if no wheel-lock was detected 

ority in deceleration performance and in wheel-lock 

resistance (i.e. yaw stability) may be considered a 

natural consequence of contemporary air brake design, 

if not equipped with wheel speed sensors of the 

anti-lock type. These unwanted characteristics deteriorate 

the control accuracy and delay the response of 

the brake system, thereby reinfbrcing its open-loop 

nature. 

Most of the transient and steady-state deviations 

(from the ideal and theoretical relationship between 

the control air pre$Iiure and the Braking Force Coefficient 

at every wheel) attenuate the brake torque. 

Therefore, insufficient deceleration performance may 

be seen as a secondary consequence of the great brake 

force variations within and between wheels. Consequently, 

it secms more appropriatc to reduce the 

brakes' sensitivity to varying operating conditions, 

and to improve the control sy$tem properties than to 

increase their peak force and to introduce mtrre valves 

or open-loop components. 

A list of sal'ety impairing characteristics particular 

tor HGVs and air brakes is given below without any 

rank order. Apart from by national governmental 

bodies, many of these problems are considered by the 

Economic Cornmission for Europe, Croup of Rapporteurs 

on Brakes and Running Gear (ECE-CRRF), bv 

the Society of Arttomotive Engineers (SAE), and by 

the International Organisation for $tandardization 

(ISO/TCZ2/SC9). An overview of the teclttrical and 

committee work issues in this context was made by 

Nordstrcim (1983, VTI report no. 257). 

a) Air brakes have substantially longer response 

times than hydraulic brakes, due to the compressibility 

and comparatively low wave propagation velocity 

of air. In addition, HGVs have long expandable air 

hoses and a long distance from the control valve at 

the brake pedal to the larthest wheel brake chamber. 

The maximum permissible response time for the worst 

brake chamber in the trailer is 0.8s according to the 

Swedish regulations (corresponding to ECE no. l3). 

b) A number of pressure modifying valves, air lines, 

and connectors may be installed by one (e.g. a trailer) 

designcr in a way that was not predicted by another 

(e.9. the axle-designer). 

c) To reduce the brake wear of their own vehicles, 

it is said that sonle owner$ install optional valves at 

the air coupling; thc truck/tractor owners enhance the 

pressure to alien trailers; and the trailer owners 

attenuate it. 

d) Many sequential brakings may consume air to 

such an extent that the spring brakes are automatically 

applied, which may cause surprising wheel-lock 

and dangerous skidding. 

e) Load sensing valves of the mechanical type 

(transforming axle suspension Inotions to pre$sure 

reductions), if mounted, are often partially or complctely 

out of order due to their tough operating 

conditions. Many types are not designed to change 

their pressure input/otttput ratio quickly enough upon 

dynarnic load transfer, which may be substantial with 

high and short vehicles. 

663

fl The brake chambers are not linear (pressure to 

force) transducers. When the diaphragm stroke and 

push-rod travel becomes long at solne wheel, the force 

declines without any easily visible indication to the 

driver. 

g) Since a brake lining may be "glazed" and lose 

much of its friction when its brake power dissipation 

is small in a number of brake applications, anti-lock 

brakes seem more favourable than load sensing valves 

even in this respect. 

h) The trailing shoe is particularly susceptible to 

glazing, since the lcading shoe makes more of the 

braking work l'rom thc beginning. Differences between 

the leading and trailing shoes may therefore be 

exaggerated, making the whole brake less efficient and 

more likely to fade or to lock up. 

i) Creat variations in temperature and friction 

coefficients for different tinings on the market add 

further balancing problems when some linings have 

been worn out. 

j) Overheating and eccentricity problems are more 

pronounced with drum brakes than with disc brakes. 

However, the peak force of air actuated disc brakes 

are often Iess than what is demanded in a HCV. 

These and other peculiarities ol HCV brakes make 

it seem very unlikely that any conventionally braked 

HGV (truck, tractor or trailer) would keep the originally 

intended brake force balance during its whole 

lifetime. Considering that many other vehicles also 

may be coupled to it, closcd-loop (anti-lock) brake 

systems appear to be essential for the braking safety 

of articulated HGVs. 

Roll Stability. Apart from increasing the overturning 

risk, the high position of the mass centre in relation 

to the track width results in great lateral load transfer 

to the outer wheels. Since the SFC decreases with 

increasing tyre load, this transfer will also aggravate 

the skidding tendency. In addition, the great lateral 

forces on the outer tyres will reduce the effective track 

to considcrably below thc nominal value. 

The roll stability becomes particularly poor it the 

load is unrestrained laterally, such as in many partially 

loaded road tankers. The lateral acceleration at 

overturning may be raised more than twice at certain 

steering f,requencies, if longitudinal cross-walls are 

mounted. See Strandberg (1978, VTI report no. 138). 

In highway speed manoeuvres the trailer is often 

moving with a greater lateral acceleration than the 

towing vehicle. Therefore, it is particularly unfortunate 

that the mass of the trailer mostly is higher 

above the road than that of the towing truck. 

Space Demarrd. The great dimensions of HGVs increase 

the probability of collisions, particularly on 

narrow roads and in urban areas. The great lcngth of 

each vehicle unit means that comparatively moderate 

sideslip angles will result in considerable lateral devia- 

664 


tions for the rear axles. Such deviations may be 

aggravated at highway speeds, if articulation or axle 

steering is introduced (to decrease the inwards offtracking 

at low speeds). 

The large front and side areas may induce dynamic 

air forces hazardous to both the unloaded HGV itself 

and to other road users. 

Indirectly Contributing Risk Factors. A number of 

design-, maintenance- or load-related factors affect 

the accident risk indirectly. For instance, the splash 

and spray from a HGV may contribute to an accident 

without any HGV involved or present. However, this 

paper will not deal further with this matter, duc to the 

diftjculties to determine the el'l'ect on safety from 

these factors and their relative importance. 

Methods to Identify Relevanl Accident Avoidance 

Parameters in HGVs. In general, deviations itt rcporting 

routines and tendency make it difficult to l'ind 

reliable numbers on travelled distance and on nonfatal 

accidents lbr valid risk comparisons between 

countries or between different vehicle types. Though 

statistics show that HCVs have considerably less 

accident risks than cars, the fatality risk (fatal crashes 

per 100 million miles) was similar to that of cars for 

single unit trucks and substantially greater for articulated 

HGVs in a U.S. study by Eicher, Robertson, 

Toth (1982, NHTSA no. HS-806-300). 

Several studies indicate that $ingle artics (tractorsemitrailer 

combinations) have better highway handling 

properties than double artics (truck and full 

trailer) and triple artics (double bottoms). Nevertheless, 

their greater low speed off-tracking may impair 

the saf'ety for unprotected road users in urban areas. 

The smaller dimensions and weight of tractorsemitrailers 

may also lead to a greater mileage in 

urban areas comparcd to truck full trailer cornbinations. 

ln addition, the yaw $tability during braking 

may be poor with semitrailers as compared to full 

trailers. In the U.K. for example, many years ago 

jacknifing occurred in about l1%o of articulated 

vehicle accidents. Load sensing valves were then fitted 

to tractor rear axles and the incidence of jacknifing 

rcduced to no more than about 5 per cent. 

Effects like this may have contributed to the higher 

accident risks for tractor-semitrailers (as compared 

with truck-trailers of similar length) found in a 

Swedish study of long vehicles by Trafiksflkerhetsutredningen 

(1977). At that time Sweden considered a 

reduction of the maximum permissible length of 

vehicle combinations from 24 to ltt metres. However, 

the Z4-metre limit appeared safer from several viewpoints 

and the l8 metre idea was abandoned. 

Though Trafiksiikerhetsutredningen compared the 

single and double artics in many way$, the poor 

matching of exposure and accident data in ol'l'icial 

statistics made it virtually impossible to isolate the

elevant parameters in the vehicles themselves. A more 

suitable method for this purpose is the case-cotrtrol 

study technique, often used in epidemiology. An 

accident group of vehicles is then compared with a 

control group passing the accident site at about the 

same time as the accident occurred. The basic idea is 

that significant group diffcrences lound in design-, 

load-, maintenance-, driver-, and employer paratneters 

reflect safety relevant factors associated with the 

vehicles themselves, since both groups have been 

exposed to the same environmental risk factors. 

Such a case-control study was conducted by Stein 

and Jones (1987, see Figure 2 above) on interstate 

highway crashes during two years in Washington 

$tate. Their results indicate clearly that certain vehicle 

parameter$, such as the number of articulartions, Inay 

be even more decisive of safety thatt driver parameters. 

This adds further doubts against the common 

conclusion that driver education is more important 

than vehicle design improvements. 

Accident Avoidance lletermined by Vehicle Design 

and Compatibility. Similar cotrclusions are often 

drawn on the basis of ambiguous results frotn accident 

investigations, stating that vehicle factors play a 

negligent causal role compared to human factors. No 

causal factor can be identified, if one does not know 

about its existence in general, and if one does not 

search for it. Many of the factors mentioncd above 

are of that kind, particularly those haeards that are 

not considered in legislation. In addition, accidents 

are multicausal phenomena by definition. 

Therefore, it is impossible to find really objective 

figures on the distribution of accident "causes" between 

drivers, vehicles and traffic environment. If 

such global cau$e categories are used, the prcsented 

figures tell more about the investigators and their 

methods than about tlte actual accidents' The impor' 

tant thing is to improve safety by the best measures 

accessible to ourselves, and to withstand the temptation 

to blame the accidents on factors that we cannot 

affect. 

This problem is particularly pronounced for HGV 

combinations, even if one accepts that vehicles are 

easier to change than basic human behaviour. The 

towing vehicle and the trailer are often designed by 

different manufacturers and sometimes they also have 

different owners. Thercl'ore, unusual efforts on the 

compatibility aspects are required by the involved 

parties to improve the accident avoidance properties 

of the whole HCV combination. 

Injury Protection 

The tables earlier in this paper show that the total 

numbers kilted per year in accidents involving lteavy 

goods vehicles varied considerably within the European 

countries, the proportions of most of the differ- 


ent categories of toad user \ilere similar. The exceptions 

were that France had a smaller proportion than 

average of pedestrian and pedal cyclist fatalities 

whereas Cermany had a higher proportion of pedal 

cyclists. 

The laboratories in Europe which have been working 

on injury protection tneasure$ for accidents involving 

Heavy Goods Vehicles are INRETS (LCB) in 

France and TRRL in the UK. A detailed investigation 

of 25 collisions between trucks and cars by INRETSa 

have confirrned the leading features o[ such incidents' 

TRRL reviewed all fatal accidents in one year in 

Great Britain involving these vehicless. These studies 

form much of the background to the following 

comments. 

In these accidents a heavy goods vehicle occupant is 

much less likely to be injured than other road users' 

Accidents that injure occupants involve either an 

HGV as a result of rollover or an HCV when it 

impacts a solid ohjcct, or when an HCV collides with 

another large vehicle. 

The largest single category of road user at risk in all 

countries in accidents involving HCVs is the car 

occupant, followed by the various unprotected road 

user.s (pedestrians and two-whccler users). 

HGV Occupant Prolection. The main causes of fatal 

injury to these occupatrts is either by ejection from 

the cab or by crushing of the cab $tructure. 

Consi

666 , 


beneficial in many of the single vehicle accidents way has the potential to increase the maximum 

where the crush forces are likely to be lower than survivable speed of impact for the car occupants. 

when other large vehicles are struck. As an example, The relative importance of special low srructures or 

if an unladen platform type HGV rolls over it is often guards at frontr sides and rear depends on several 

the cab roof that collapses and crushes the occupant. factors. It is difficult to justify sideguards fitted to 

More substantial pillars connecting the roof to the rest trucks to protect car occupants. They would of 

of the cab together with more secure glazing might necessity have to be fitted to both sides of trucks and, 

also prevent the windscreen coming out and therefore because of their length and strength requirements, 

reduce the incidence of ejection. However strong the would be very heavy. Because of the weight and 

cab structure, it is difficult to envisage much protec- payload penalties incurred and the relatively small 

tion being offered in the far more violent HCV to benefit in terms of lives and injuries prevented, they 

HCV type of accident or when the HGV impacts a would probably not be cost effective. 

very solid object such as a bridge parapet or other Guards and special low structures fitted to the front 

roadside furniture. However the first design of the and rear of the heavy goods vehicle are more likely to 

Leyland National bus did have the structure around be cost effective because of their lighter weight and 

the driver locally strengrhened $o that it displaced their potential, at least as far as the front guard is 

backwards without crushing the driver in frontal coilcerned, to save more lives. The mechanism of 

impacts. 

impact of cars into the fronts and rears of trucks is 

At present only Germany and Sweden have require- similar except that there is usually a greater degree of 

ments for cab strength. 

under-run at the rear of the goods vehicle because of 

its high structure and space under the rear. Also the 

Car Occupant Protection. Heavy goods vehicles are speeds of impact of cars into the fronts of trucks are 

very aggressive towards cars in collisions. The large usually higher. Because of the much larger number of 

mass ratio ensures that the velocity change of the car car occupants killed in impacts into the fronts of 

is much greater than that of the truck. The height of trucks compared with those killed in impacts into the 

structure around the perimeter of most trucks is such rear, there is a stronger case for the fitment of 

that when it $trikes or is struck by a car, the car can protection for cars at the front. 

under-run the truck, often to the extent that the truck A British study in 19857 suggested that an estimared 

structure comes into direct contact with the car 60 car occupants might be saved out of about 2.000 

occupant. By the same mechanism the importarrt killed each year in Creat Britain. This was based on 

energy absorbing zones of the car, which tend to be fitting energy absorbing front under-run guards and 

below the truck structure, are not used to their be$t an experimental impact test programme suggested that 

advantage. Finally, the rigidity of the truck structure such guards could offer protection to seat belted car 

ensures that most of the energy of the impact is occupants at closing speeds up to 65 km/h 

dissipated in the car structure rather than in that of The concept of including energy absorption 

the truck. 

in 

guard design is important, particularly in front of 

Typically, in European countries, the distribution of truck to front of car impacts where closing speeds are 

impact of cars around the truck perimeter is that higher. There is however probably a limit in rhe 

approximately 60 per cent or more impact into the amount of energy absorption that may be provided. 

front of the truck (usually front of car to front of The linear crush of the car plus the crush of the guard 

truck), around 25 per cent or more into the sides and must not exceed the original length of the bonnet of 

up to 15 per cent into the rear. 

the car otherwise the upper structure of the truck may 

In all these types of impact, the important primary impinge on the car occupant compartment. Also the 

objective is to provide a strong structure or guard, forces necessary to deform the truck guard must lie 

fitted to the heavy goods vehicle, which is low enough within the range that will also deform the car srruc- 

to impact the car structure. Two advantages are then ture. These two factors may well imply that the design 

gained by the car occupant-the truck structure is of the low front or truck guard is closely determined 

kept away from the car passenger compartment and by the dimensions and crush forces of small cars. The 

the energy absorbing properties of the car arc utilised. previously mentioned British study showed that for 

The latter is only a real advantage if the car occupant small car impacts it was possible to utilise at least Zj 

is wearing a seat belt. A secondary but important kJ of energy absorption built into a guard., 

desirable objective is to make the guard or structure The ground clearanoe of the guards is also impor_ 

which strikes the car encrgy absorbing. For this to be tant and would need to be about 300 mm for good 

effective the forces needed to deform the guard perforrnance, and certainly no more than 4{X) mm, 

should be compatible with those required ro deform otherwise the structural parts of smaller cars would be 

the car structure. Energy absorption introduced in this overridden.

As yet there are no European countries that requirc 

the fitment of front protection to heavy goods vehicles. 

Many countries do however have a requirement 

to fit rear protection, mainly to meet EEC Directive 

70/221. These countries include Austria, Belgium, 

Great Britain, West Germany, ltaly, Luxembourg and 

the Netherlands. Sweden has also had a mandatory 

requirement to fit rear guards, to their own specification, 

since 1973. France also has a national requirement. 

It has not yet been possible to estimate what its 

effectiveness has been, but a large improvement is not 

expected because the incidence of cars striking the 

rears of trucks seems to have been declining over the 

year$. 

With the increased use, throughout Europe, of seat 

belts in cars, the case for fitting trucks with both 

front and rear protection is now much stronger. 

Considerable research has been done and is being 

done in countries such as Germany, Sweden, France 

and Creat Britain to develop and promote effective 

structures and guards. 

ln France in 1985. the "Laboratoire des Chocs et 

de Biomecanique" of INRETS started a new experimental 

research programme on compatibility between 

cars and HCVs in frontal impact. The first phase 

consists in determining the geometry of an "avergggtt 

HGV. This work based on the sizes of 20 trucks is 

completed and suggests the dimensions of a rigid 

barrier to simulate a truck front end. In the second 

phase which is in progress several cars will be tested 

against this obstacle at two impact speeds: 50 and 57 

km/h. Already one car model has been tested at the 

two speeds. The first results seem to show that the 

injury parameters currently used to evaluate the protection 

of car occupants are not sufficient, and are 

not really valid for such a crash configuration. The 

third phase will lead to the proposal of a methodology 

for the evaluation of compatibility between trucks and 

cars. 

ln the Federal Republic of Germany the Technical 

University of Berlin has investigated the performance 

of energy absorbing low front bumpers to protect cars 

against under-run. 

In Great Britain the TRRL is developing a test 

procedure for evaluating frontal impact protection on 

trucks and heavy vehicles. It consists of a set of 

dynamic impact tests using an impactor of 250 kg 

which strikes the truck front at several different 

points acrosri the width at low bumper height" 

Pedestrian, Pedal Cyclist flnd Motorcyclist Protection. 

The unprotected road user, as this group is collectively 

known, is mainly at risk from being run over at 

the sides of the trucks. The cutting-in of articulated 

vehicles as they turn sharply can trap pedestrians and 

others. ln the Netherlands a down looking nearside 

mirror is required which may help drivers avoid this. 


Pedal cyclists arc also at risk in normal overtaking 

manoeuvers when they may topple towards the vehicle 

as it overtakes them and fall under the side and be 

run over by the wheels. 

Lightweight sideguards would help prevent running 

over but their design should be considered carefully in 

order to make them effective. 

A large proportion of accidents where pedestrians 

or two-wheeler users fall into the path of the truck 

wheels occur, not surprisingly, in urban areas and 

often at quite low speeds. Guards need to be strong 

enough to withstand normal everyday use but from an 

accident point of view, need not be very strong. More 

important requirements are that, if the guard is a 

horizontal rail type of structure, the rails should be 

close enough to prevent the road users falling through 

them and the supporting structure should be recessed 

to prevent injury. The whole guard should also be 

close to the outside edge of the vehicle or trailer to 

reduce head injury caused by the aggressive protrusions 

such as loading hooks often to be found in that 

area. Ground clearance is possibly the most important 

factor and it should be as low as operating conditions 

will allow. A maximum height of the lowest edge of 

the guard from the ground of about 400 mm would 

minimise the tendency of a road user to roll under the 

guard. 

All European countries have similar problems in 

protecting their pedestrians and two-wheeler users and 

should benefit from the fitting of sideguards to 

HOVs. At present only France, the Netherlands and 

Great Britain have a requirement for them to be 

fitted. The subject of sideguards is however being 

considered by ECE Geneva with the intention of 

producing an ECE Regulation to specify their design 

and performance. Their potential for saving lives will 

of cour$e vary slightly from country to country but as 

an example it has been estimated that in Creat Britain 

about 50 lives per year could be saved by fitting 

effective sideguards, but less stringent requirements 

could lead to this being more than halved. 

Priority of Measures to Provide Injury Protection, 

The following is suggested for the order in which 

measures to provide injury prevention should be 

introduced. It is based only on the liklihood of saving 

lives rather than on cost benefit considerations and it 

assumes that nearly all car occupants wear their seat 

belts. 

l. Frant under-run messures (guards or low 

struclure) 

Energy ab$orption should be incorporated 

and ground clearance should be around 300 

mm. 

2. Sideguards 

Lightweight structures are needed with a 

ground clearance of no more than 400 mm 

667

and several other detailed but minor requirements 

if their effectiveness is not to be 

needlessly reduced. 

3. HGV occupant prolection 

Seat belts suitable for trucks would be a 

very cost effective means of' preventing ejection 

if they were to be worn by drivers. Air 

bags might be an altertrative but tnight be 

less effective in some ejection situations. 

Stronger cab structures are also desirable to 

reduce the injuries due to crushing and 

would also make seat belts more worthwhile. 

4. Rear under-run guards 

The priority here is less than for front 

guards but similar requirements are needed. 

The primary need is to prevent under-run but 

energy absorption can also be beneficial. 

Methods of Introducing Injury Protection. With the 

exception of HGV occupant protection, the measures 

to protect other road u$ers are of little benetit to the 

truck operator. He has to pay for the devices. The 

only economic advantages lie in the possible reduction 

of aerodynamic drag by suitable design of the guards 

fitted to the tiont and sides of the truck. For example 

the front guard could incorporate an air dam8 and thc 

sideguards could be of skinned dcsign. Both these 

ideas have been shown to reduce drag and also to 

reduce spray around the vehicle. The only other 

advantage is that front and rear guards rnay reduce 

the extent of damage to the truck itselt. 

These factors may not be sufficient to persuade 

many operators to fit protection, in which case 

compulsory fitment through legislation is the only 

certain way to improve the situation. This method is 

never popular with operators or mflnufacturers even if 

there is a strong case for the proposed legislation. It 

may be possible occasionally to trade ofl benefits to 

the operators in exchange for enhanced safety. For 

example, in Great Britain in 1984, sidcguards and rear 

under-run guards were introduced at the sailc time as 

an increase in the permissible maximum gross weight. 

Although this eased the introduction of safety features 

there was opposition from many sources to the 

increased weight. 

Conclusions 

The last ten years has seen a continuation of the 

improvements in the accident situation for Heavy 

Goods Vehicles in most European countries which had 

been apparent during the previous ten years' Frogress 

in reductions in casualties has been made in different 

countries at different times in this period and not 

always for reasons that are readily apparent, To some 

degree these results are a consequence of changes in 

the transport of goods. In some countries the numbers 

of goods vehicles has increased while in others it has 

668 


decreased, Total distances travelled have been fairly 

stable but there has been a tendency to increase. It is 

clear that in many countries there has been a tendency 

towards higher capacity vehicles. As a result the total 

of goods transported has probably increased. Generally 

increases in goods traffic have been greater in 

European countries outside the EEVC membership. 

Precise comparisons in accident trends between 

countries are difficult to determine because of differences 

in the minimum size of vehicle regarded as 

being a Heavy Good Vehicle. However even between 

the four courltries studied in detail (France, Federal 

Republic o[ Germany, Sweden and Great Britain) 

there is a factor of at least two to one between the 

highest and the lowest thtal casualty rate measured 

against the various base statistics. The rea$ons are 

clearly a combination of geographic and dcmographic 

factors combined with goods vehicle factors such as 

the amourrts of goods transported by road, the road 

user behaviour prevalent in thc different countries. 

The lowcst national fatal accident rate per million km 

travelled by HGVs (l and 2 vehicle accidents only) is 

about 0.04. 

In the paper the fatal casualties in HCV accidents 

are sub-divided according to their road user category. 

About l09o are occupants of the HGVs, while car 

occupants make up between 50o/o and 65Vo and 

pedestrians about 1590 (although only 890 in France). 

The remainder are mostly riders of two wheelers. 

These figures suggest that there are five main factors 

determining the accident rates for these heavy vehicles- 

Road (desien and conditions). 

I HGV (design and operation and divided 

between accident avoidance and protective 

features). 

I Other vehicles involved (design and operation 

and divided between accident avoidance and 

protective features). 

I HCV drivers (skill, behaviour and operational 

conditions). 

o Other road users involved (speed and behaviour)' 

This report deals only with HCV vehicle factors from 

among these five. 

The review of accident avoidance possibilities for 

HCVs starts with a list of institutions interested in 

their study. lt generally concludes that there is a large 

diversity of design practice and requirements in Europc, 

never-the-less progress in this atipect of safety is 

readily possible. 

The stability in yaw of the multi axle HCV is 

shown to depend on the load balance between its 

many wheels in relation to the side force demands 

made on them when cornering and in conditions of

Iow grip. The problems increase a$ the numbers of 

articulatisns and axles are increased. 

These problerns are compounded when braking is 

involved. The brake distribution has to be set and the 

maximum braking available without locking wheels is 

dependant on many factors which are listed. Various 

systems of valves to alleviate the situation are mentioned 

as are the limitations of current air brake 

system$. It may be concluded that anti locking brakes 

and electronic control of the braking of each wheel 

are essential if full braking with stability is to be 

achieved. 

Limited stability can lead to an HGV not always 

following in the path taken by its driver in ditticult 

conditions and this leads to accidents, especially on 

narrow and on crowded roads. 

Roll stability is limiting when the load is placed 

high and also for fully and partially loaded tankers. 

Lowering the height of the load is perhaps the only 

viable design improvement. 

The review of injury protection possibilities for 

road users involved in accidents with HCVs showed 

that for their occupants the most desirable measures 

are those to prevent ejection and crushing within the 

cab. Seat belts might prevent 3090 of all fatalities if 

worn. Several countries require strengthening of cab 

structure$ in overturning but the great need is to 

better resist longitudinal crushing on to the driver in 

frontal impacts. 

Car occupants whose cars strike HGVs can best be 

protected by front and to a lesser extent rear underrun 

bumpers or low structures. Studies are well in hand 

for front underrun protection, while rear bumpers are 

already required. 

Unprotected road u$ers can best be helped by the 

provision of'side guards to prevent them falling under 

the sides of HGVs and then being run over. Several 

European countries require them to be fitted and the 

ECE Geneva organisation is working towards a Regulation. 

As for the underrun protective features the 

overall effectiveness is much increased by careful 

detailed design of the guards. 

It is concluded that the design of Heavy Goods 

Vehicles can further be improved in a number of 

different ways to reduce its contribution to the road 

casualty situation. Most of these features are being 

studied at either the research or the design stages. 


References 

l. Hotop, R. Lkw-Geschwindigkeiten auf den Bundesautobahnen. 

Stradenverkehrstechnik, Heft 

5/1985. Kirschbaum-Verlag, Bonn-Bad Godesberg. 

Deutsches Institut fur Wirtschaftsforschung 

(DIW): "Verkehr in Zahlen 1985" Hrsg.: BMV, 

Bonn, September 1985, 

Statistisches Bundesamt Wiesbaden: StraBenverk" 

ehrsunfhlle 1984 Fachserie 8. Reihe 3.3 Stuttcart/ 

Mainz 1985. 

4. Dejeammes, M. Heavy trucks aggressivity for 

road users-in search of improved safety. Proceedings 

of the lOth International Technical Conference 

on Experimental Safety Vehicles, Oxford 

1985. NHTSA US Department of Transportation 

5. Riley, B S and H J Bates. Fatal Accidents in 

Creat Britain in 1976 involvirtg heavy goods 

vehicles. Department of the Environment Department 

of Transport, TRRL Supplementary Report 

SR 586. Crowthorne l9B0 (Transport and Road 

Research Laboratory). 

Riley, B S, Chinn, B P and H J Bates. An 

analysis of fatalities in heavy goods vehicle accidents. 

Department of the Environment Department 

of Transport, TRRL Report LR 1033, 

Crowthorne l98l (Transport and Road Research 

Laboratory). 

Riley, B S, Penoyre, S and H J Bates. Protecting 

car occupants, pedestrians and cyclists in accidents 

involving heavy goods vehicles by using 

front underrun bumpers and sideguards. Proceedings 

of l0th International Technical Conference 

on Experimental Safety Vehicles, Oxford 1985. 

NHTSA US Department of Transportation. 

Tromp, J P M. Splash and spray by lorries.Institute 

lor Road Safety Research SWOV, The Netherlands. 

Report R-85-5, Leidschendam 1985. 

Strandberg, L. "On 2. 

3. 

6. 

7. 

8. 

9. 

the braking safety of articulated 

heavy freight vehicles." Vol. 2 of Proceedings 

Sympasium on the Role of Heavy fr'reight 

Vehicles in Traffic Accidents, Montreal 1987. 

Transport Canada, Ottawa, KIA, ON5. 

Twelvc additional ret'erences are given directly in 

the text on Safety Measures and on Accident 

Avoidance. 

669

h 

qJ 

g) 

E) 

fiE 

o 

F l O 

r(b 

.l tH 

E h 

I = 

(l).= 

ELH 

r = 

6 

z 

t 

F 

.E 

I E 

T' 

It 

t 

o 

an 

+, 

6 o 

t 

o 

(, 

E 

s 

L A 

ott 

fiE 

otr 

h t 

r-{ l) 

gtr 

El '.{ 

H 

tt 

F 

to 

+, 

s 

E0|, 

.dcut 

troE 

F n . ! 

t{+1 () 

GlFI 

nF 

liFl O 

Q o b 0 

+r 'Fa O 

*6fi 

tst) 

n 

F 

s 

F 

t, 

I 

T 

u: 

F 

! 

F 

D F 

,3 


It{ 

+ J . 

t8 

P *F 

.{ -{l 

t t l 

3 E 0 

O Ei t\,| 

f, '{ r{ 

5.lt or 

.olrE 

BF fif, 

d l l ) 

3EE$ 

EE.c.c 

-tr +r +t 

t! tll u0 E 

FiO C tr 

5d.3"3 

I . 

8$ 

hB i ,i 

flEgFFI 

E iq E E E 

F Ed ; ; E 

i :8 i d B 

I $B I I I 

s .HE s s s 

rD 

C'| BB 8 3 3 3 3 

(\| loto 

dt6l 1 8 I I I 

N 

F{ 

H 

C) 

HI 

;l 

Fq 

f;l 

HI 

o.l 

I 

F 

I 

t) 

(J 

t t) 

6 

(J 

E 

(\|lD 

F{dl 

F{ 

( \ | l o e 

' { ( l | g l 

rO 

gt 

rO 

ol 

( q ( \ | A A 

u*=fi rH rH rH 

.$ H6 Hd Hd Ed 

E O 

o o 

H H 

EI N 

0tl 

(\| 

Fll 

{JI E E 

vi t v{ 'r{ 

Êl b0 h0 

[tl Y{ 

}Itr 

Yl 

E 

r r 

0t 

O 

D tt O TtF{ 

o t l ) o x 

Fl +r-l 4) 6 

F lx lux ql 

d F{l!t F{O 

j A 5 L 

.ltt tJrt o O 

tt v{ v{l 

bDE +JE +, 

EE IE +b 

EO 

E 

# u 

r L 

fl +r 

L E l 

g't 

$ " 

6 t { 

f l f 

l L d 

E ! d 

t r l 

Y{ l! 

F o 

F \ E 

o 6 r 

h nE 

rH q .'{ 

6sl 

8 0 0 

C

l(rtronal regulrtion tor tha optrdtion of ythlclCt 

styz0 

lhtor vrhlrlr trtcgofy 


Appendix f ft) 

cfipAfltsto( oF ttAIlottAt ff(o IirtRr(AIl0lrAl Rt6utlIl0lt5 FoR tr{t oPtMII0t 0F vtHtctts 

FtotML RtPUEI lc 0F GERHATiY, stvzo'r trttl'to tctt 

0ri vi n9- 

I it?nce 

clrss 

5perd. 

l-i mi t\ 

!./li 

Grot 3 

vrhiclt 

rrr 9ht r 

Fr tug 

ntrrnatidnrl rrgulltion rorrrsponding to 

llotor 

vehiclr 

ca trgory 

rtyin9 

I tEnft 

lrsco 

ttc rnd Ect 

Rodd Irrffir R?gistrdtidn AEt 

' 

Direttions of the [uropesn tconoilit CoimJnity for rotd Yth{tItr (tfC-dfrections} 

t"d!' 

Prr:rngtt crrt t too ( ) 7.5 rl I t.5 

ilotbr buts?s 

t 

? lxlct 

e rrlas 

? trndr* rrltt 

Articultttd bstrci 

Htrvy goodt vchltlrt 

? rr lc: 

2 rxll5 

3 rrll5 

I rrlts 

Artirulrted Ythicle3 

llucters ot the rrles 

Semi-trrilrr seFi _trrllerg 

trucki 

z l 

? ? 

? 3 

3 1 

lrtlculrted Yehirlrr yith 

I 50 -tontr i ner 

Trrilrrrr 

1 2 

l 3 

I rrlP 

? rrles 

? arles 

I rrles 

2 

2 

2 

? 

3 

? 

2 

2 

t 

? 

? 

a 

? 

z 

3 

2 

? 

t 

80 (Eo) 

80 (80) 

B0 (80) 

80 (80) 

E0 (80) 

60 (80) 

60 {Bo) 

60 (80) 

60 (80) 

6o {Bo) 

60 (80) 

60 (80) 

60 (80) 

60 (80) 

80 (80) 

60 (80) 

60 (80) 

60 (80) 

J 

Frgulrtlont of the tconomic Comisrion for Europt for mtor vthicles lnd their trailerg 

t 

0ut of twns, ( ) on highH.ys (Autobrhh) 

t 

5"o.rlt. srttl.nent for frontlrr trdffit in tht slrrlrnd (l Y. StYfO) 

sinqle rrle 

llt 

trnJem r'lts, rxlr bit? Ll,35n ?lt 

t6 

zt 

30 

?8 

2,8-l ,3 

It 

2l 

x2 

2l 

tl 

t0 

r0 

t{ 

l0 

t6 

7a 

#il: I:l;:l: :: l::i;:: l;ll l.ii'Ii." , ,,'., lll 

rctor vehicle3 erth I rrlcl rnd trlrler Fith ? trlrs 3gt 

r 

ltt drtr refer to the intenationdl driving lltencc 

I 

Yith porrenge', tithout re.tsr roxirum speed: 60 (60) 

I g.rving licrncr ct.tl t is necesrrry for truin vith mrf tlrra ! iilcr. vithDut rrgrrd to thG pulllng vehitlt 

H? 

nl 

t{l 

t{? 

n3 

0l 

n2 

03 

0r 

D 

0 

E 

c 

t 

rJi 

>5 

ig t.5 

>l.5El? 

-l? 

=o.rs 

i 

E0.7F-r.5: 

| -r.=to ' 

I -ro 

Srr tt 

>E 

>0 

Ixplrnrt{ont 

0n highxryl (rutobahn) 

no Jpeed lifiit. out .d. 

f LLqd -LPggd -ol__ll0 ri{!- 

IDD ln/h if 

-the rux. speed 100 \h/h 

(it depends on th. ca. 

I i rence ) 

-the rnginr pffrr ll tl/t 

of the grogt Yahltlt 

reight 

-"100" brdgE rith serl 

Irucls up to ?.8t rrr 

trertrd llle prri?ngtf 

ctr9 

Driving licrn(e'E'li 

n€ce3srry tor oparrtlofi 

rlth r trtiltr 

Grosg Yehirlr reight 

rrting by; 

firsl registrrtion 

rfter l9.ol.'87 ;lst 

fron dEC.'9t rll :35t 

Iu3t not ruFp,t! 

th? 9ro33 vehicle reight 

qf the 0ullrng tttor 

vEhirlr

Appendix 2 (a) 

Improvement of road situation in Europe 

Great Britain 

Road Lengths in 1000 km. 

Yetr fotal llotorray ilon-bul lt-up 

t973 

l97a 

t9?5 

1976 

ts77 

ls78 

1979 

lgao 

l96l 

l9B2 

1983 

r9a4 

l9A5 

327 -l 

329,O 

330,O 

333,4 

336.7 

336.3 

33a.o 

339,6 

3dt.9 

343.6 

3d5,4 

347,2 

3r$.3 

1.731 

1.a69 

1.97s 

2.155 

2.236 

2.394 

2.455 

?.s56 

2.62A 

2.666 

?..7m 

P.802 

?,834 

Appendix 2 (h) 

Improvement of Road Situation in Europe 

Germany 

19s.5 

t99.O 

t99.e 

200.3 

200. I 

2()0.5 

eol,o 

20().9 

20r.2 

20r,9 

2oz. I 

202.8 

?o2.6 

l. Foada rlth lpaed ltrlta of 64 kfr/h oF lcae. 

Yer r 

| 970 

t 97l 

1972 

roTt 

1974 

1975 

1976 

1977 

I 978 

197 9 

| 980 

l 98l 

| 98? 

I 983 

| 98{ 

r 98s 

rotal 

r6?,3 

r64,5 

t65,3 

r66,i 

t67,5 

t 68,2 

t69,1 

r69,6 

r70,,| 

t 70,7 

t7t,5 

t7?,4 

t72,5 

1 73,0 

t7?,6 

t73,0 

Rodd [ength In lDoD tm 

Road Crtegory 

outside build u 

I 

rist'*or 

4 ,110 

4,461 

4,8?8 

5,?58 

5.48 1 

5 ,748 

6.?13 

6,435 

6,7 il 

7 ,029 

7 ,292 

7,538 

7,784 

7,919 

I,080 

8, 198 

EXFERIMENTAL SAFETY VEHICLES 

: 

t 

Butlt-up 

r26,9 

t28.2 

t28.e 

l3l.o 

txz.4 

133,3 

t34.5 

136.2 

l3a. I 

t59.O 

ldo,5 

l{1.6 

142.9 

trPtS 

Rurrl urb! h 

t 58,3 

r60,0 

160,5 

161,{ 

r6?,0 

162,4 

r63,0 

163,1 

t63,3 

163,7 

t64,2 

t64,9 

t64,7 

t65,0 

r64,6 

t64,9 

270 

276 

28? 

286 

?90 

294 

?99 

30? 

3Ds 

308 

310 

312 

314 

316 

317 

Lengths of public roads according to road 

categories 

/ 2 / 

Hidth of 

the roads 

l. l. 1966 

snaller { n 

4- 5nr 

5 - 6 n 

6 - I n 

7 - 9n 

9-l?rn 

l? rnd nore 

tota I 

r. r. r97 l 

smaller 4 rn 

4- 5nr 

5- 6nr 

6 - I n 

7 - 9nr 

9- r?m 

l? and more 

tota I 

Jg 

smaller 4 m 

4 - 5 m 

5 - 6 n 

6 - 7 m 

l - 9m 

e-t?m 

l? and more 

tota I 

1.L t9Bl 

smal ler 4 nr 

4- 5rn 

5- 6nr 

6 - 7 n 

7 - 9 n 

9 - 1 ? m 

l? and more 

totnl 

Appendix 2 (b) 

continued 

Road category 

H{ghh,ayl Rurrl I 

(Autobahn)l 

I 

?6 

3296 

337? 

69 

4 39? 

446 | 

80 

6127 

621 3 

135 

7402 

7538 

il 5?7 

395 78 

5?449 

30238 

r 5980 

u6t5 

t 798 

t54160 

7848 

3r557 

5207? 

3895? 

?t908 

3913 

u991 

r60008 

536 | 

t5ttu 

53r97 

444 

t4 

?7495 

5375 

38?3 

t62933 

4?06 

20??7 

| 9l5B 

48038 

3?t9l 

638? 

4657 

U"brn 

97t88 

75503 

42678 

I 9356 

I 5385 

2502 | 9 

| 0209? 

77627 

52888 

250 t6 

18752 

?76375 

t0z3f7 

795!0 

6?58s 

3uo37 

?t699 

2 96 73e 

l5q 854 3 I 0000 

Lengths of public roads according to 

road widths and categories 

/ ? /

SECTTON 4. TECHNICAL SESSIONS 

Appendix 3 (a) 

Heavy Goods Vehicles and Trailers* 

Percentage failure by category in Great Britain**: 

Body t Chassis 

Suspens I on 

Erhaust 

Tyre5 E llheels 

5teer i n9 

BrBkes 

Lights 

ruel t Tanl 

Tachograph 

0il & lldste 

Ilectrlcrl 

0thers 

t979l80 1983/84 r984/85 r985/86 

4.51 X 

8.01 r 

?.17 x 

5,4 I 

3. I X 

45.97 X 

9.17 X 

'.: ' 

1.47 Z 

0.96 X 

8. 13 x 

4,89 X 

8.t5 X 

?,35 r 

6.39 I 

2.89 z 

48,15 I 

11.59 I 

1.33 X 

3.5 I 

1,72 I 

0.9t x 

9.03 Z 

Appendix 3 (b) 

4.63 I 

7,83 X 

2.t8 t 

?.77 \ 

49.14 N 

11.9 I 

1.33 I 

?.71 Z 

l.5l I 

0.89 I 

9.52 z 

4.49 r 

8.07 r 

2.01 | 

6.75 | 

2.83 I 

51 .26 1 

12.55 x 

1.38 x 

2.49 7 

1.44 I 

0.9? | 

10.12 z 

Total numbsrs of dcfccta nlqt 90 1647 girllirlil !64340 

. Goodr vehicles over 1525 kg unlrden reight rnd tr8ilers over 

Io2o ka 6lsd€n s€iaht. 

il The table shdvs the percentdge of vehicl?s failing as I result ot 

the defect indicated, A vehitle my have more thlh one fiult and 

consequently the t0tals o+ the percentdges 0f separdte faults 

exceed the percent6ge of vehicles falled. 

Results of generat technical inspections for motor vehicles according to type of defect for Germany(3) 

letr 

I 970 

| 975 

| 980 

l98 l 

198? 

| 983 

| 9{r4 

| 970 

I 975 

| 980 

l 98l 

r 98? 

| 993 

r9M 

l 970 

| 975 

r 980 

t98l 

| 982 

t 983 

r 984 

lel I tnE stock 

fl6-tdFTerrr'l?r 

rnd 

Trr i I ers 

in l't000-i 

t7dl5 

?19{? 

?8?67 

?9077 

?9664 

3034 2 

3l l6? 

Crrl rnd lfrqons 

13941 

| 7898 

?319? 

?3730 

?n 105 

?4 580 

?5? 1g 

lruckg 

il19 

t?45 

l4it8 

t168 

tit 7g 

t4 75 

I lt87 

lrlentlfled Defects 

l-rn l00o-7 

ro*, 

I rrsr,t, 

| ::*;u | 0.,*.* 

730 7 

| 0955 

t004 | 

t0533 

10498 

r0r50 

t0238 

s739 

9?s5 

84 90 

8989 

89?3 

8644 

968? 

790 

66f 

517 

540 

507 

16? 

145 

| 360 

?r3l 

?0s0 

? r64 

?l 1g 

?0?9 

|99t 

t061 

t805 

t 770 

r85? 

I 795 

1720 

t6 70 

t?6 

l14 

r0l 

99 

93 

86 

83 

86t 

782 

7l I 

681 

684 

668 

694 

l16 

55s 

603 

978 

978 

559 

588 

73 

4? 

36 

3? 

30 

27 

?5 

166 5 

?460 

I 996 

?24? 

2 r95 

? ll'9 

?0 75 

| 358 

?r58 

| 754 

| 985 

| 938 

186 I 

| 8?0 

t5B 

t43 

99 

t06 

9B 

89 

B6 

| rr*' I,::lii::"" | ,::nil 

503 

536 

423 

42? 

q4g 

452 

453 

39s 

430 

319 

3?0 

340 

345 

354 

54 

38 

26 

?4 

?3 

?? 

?1 

l 009 

?178 

??86 

?558 

?599 

?64 I 

27 74 

t$ 

1797 

r 93? 

?20 9 

??4 0 

??90 

?395 

r39 

t53 

r3l 

t34 

t?5 

ilB 

t l5 

383 

851 

833 

8lt 

798 

770 

786 

3?8 

798 

783 

762 

745 

t7l 

732 

3? 

3l 

?6 

?5 

24 

T( 

It


Priorities in the Active and Passive Safetyof 

Trucks 

K. Langwieder, 

M. Danner. 

HUK-Verband 

Automobile Engineering Department 

German Association of Third Party, 

Accident, 

Motor Vehicle and Legal Protection Insurers 

Federal Republic of Germany 

Abstract 

In a representative study of some 1,2(X) accidents 

the accident characteristics of the truck were ascertained 

and priorities for partner-protection measures 

were derived from them. 

A truck front protection in collisions with cars and 

side guards for pedestrians and motorcyclists are of 

paramount importance. The protective effect deduced 

by the rear underride protection hitherto and the 

necessity for further improvements is proved from the 

accident analyses. 

So far only a few studies are available on the 

accident risks for the truck driver himself . In a 

representative evaluation of accidents the injury patterns 

of truck drivers dependent on collision type and 

accident intensity are shown on the basis of 800 

truck-collisions. This is followed by a discussion of 

the injury causes with regard to possible safety 

improvements. The active safety requirements of 

trucks resulting from accident studies and the possible 

effect of technical sy$tems such as antilock braking 

system$, distance warning devices etc., are discussed. 

674 

\ffl"J*,,,,,", 

Accidents 

with 

fatali t ies 

Accidents in 

Germany l9t5 7 678 t 

Accidents involving 

trucks dnd 

pcrsonal iniuryr 

1 385 

Accidents with 

seriws iniuries 

In the studies the distinction was also made between 

the various truck categories. 

The Present Situation 

Truck accidents constitute an important element in 

the general accident scene. Each year in Germany 

about 30,000 accidents with personal injury involving 

trucks occur /1/, in which some 1,700 people are 

killed and 40,000 are injured (Figure l). In addition, 

in 1985, about 36,000 truck accidents occurred causing 

serious material damage. Although in view of 

their mileage the involvement of trucks cannot be 

described as overproportionate, more attention in the 

future must be paid to truck accidents if the absolute 

accident figures are taken into account, which frequently 

not only involve extremely serious injuries but 

also an extraordinarily hieh level of material damage 

and a considerable risk to the other traffic in general. 

It is chiefly the accident opponent of trucks who 

run the greatest risk of injury. For about every two 

carlcar-collisions with fatalities there was one 

truck/car-collision with fatal injuries (Figure 2). Compared 

with the safety efforts up to now to reduce the 

risks in car/car-collisions future truck safety research 

must be considerably intensified in view of this fact. 

But neither should the efforts to achieve more protection 

flor other road users like motorcyclists, pedestrians, 

and more safety for the truck occupants themselves 

be relaxed: even if the risks to the truck 

occupants are not high in relation to the trucks 

registered, the absolute injury figures are not insignificant 

and the potential safety reserves must also be 

used for the truck occupants. 

Accidenti with 

sliBht iniuriei 

Accidents with 

injur ics/farrl i ties 

Total 

320 067 327 7t-5 

oc40 r8 532 28 163 

of these iniured 

truck trcupants 12s I 803 5727 7659 

rfurtlrr 

tfuck accid€nts with serious material damage in lgtft lI 924 cases 

Figure 1. Truck accldents in Germany in 1985 wlth resultlng Inlurles to occupants

I 


7 678 *.io."tt 

wrth latdlrrr(r 

I400 f{r{rirics 

Flgure 2. Trafflc accldente ln Germany in 1985 wlth fatalities 

In view of the existing accident risks, it is astonishing 

that up to now only relatively few studies have 

dealt with truck accidents, emphasis usually being 

placed on the consequences of the accidents and only 

rarely on the general accident characteristics and the 

crisis situations in the pre-crash-phase. 

It is the aim of this paper to give an overview of 

the accident involvement of trucks and the findings 

available by the HUK research on truck accidents with 

respect to 

I 

t 

I 

partner protection in truck accidents 

self-protection of the truck occupants and 

active saf'ety of trucks 

Methodical Aspects of Truck Accident 

Studies 

Building up truck accident rnaterial involves extrernely 

great problems, since truck accidents are more 

difficult to collect by occurrence, call for supraregional 

evaluation because of their high percentage on 

highways and since the evaluated accident material 

has to be divided into a large nurttber of differing 

truck concepts. 

Further problems result from the fact that-unlike 

in the case of car-or motorcycle-accidents-there is 

no clear reference basis I'or critcria of representativity 

of the material. 

The definition of the term "Lastkraftwagen" 

(truck) as given by the Federal German Statistical 

Office covers all vehicles which are suitable for 

transporting goods. It thus necessarily includes a large 

number of completely different vehicle types-vans 

and trucks with a tenfold variation of mass and 

fundarnentally differing concept and use. 

For this reason only trucks of over 3.5 tonnes 

maximum laden weight with typical tmck construction 

were considered in this HUK accident material presented; 

moreover, a detailed classification by mass 

categories and the type of truck, trailer and articulated 

lorries was carried out. 

Two independent bodies of accident material were 

available: 

' 

I Truck material I 

an evaluation of 1,447 truck accidents with 

car$, cycles, motorcycles and pedestrians 

with regard to problems of partner protection. 

a Truck material II 

an analysis of 770 accidents with truck/ 

truck- and truck/car participation and single 

truck accidents with regard to the problem 

on self protection for the truck occupants 

and of active safety. With regard to ques- 

' 

tions of active safety, 300 of these accidents-a 

representative sarnple of tmck accidents 

in the Federal State of Bavaria in 

, 1984*were analysed and examined in-depth. 

For these studies which were partly conducted in 

cooperation with the German "Forschungsgemeinschaft 

Automobiltechnik" (FAT) and thus with the 

truck manufacturers well over 10.000 accidents had to 

be evaluated in order to filter out this accident 

material. 

675

The accident material now available will be continuously 

expanded in the next few years to create an 

enlarged basis for special studies, for example, in the 

case of truck categories or accidents involving dangerous 

goods, and to examine the effects of typical safety 

constructions. 

Passive Safety of Trucks-Partner 

Protection 

Car/truck collisions must be given prominence in 

measures of partner protection (Figure 2), although 

truck collisions with pedestrians and motorcyclists 

also represent a high proportion of accidents with 

fatalities. Although ways of protecring cyclists, motorcyclists 

and pedestrians in the case of truck collisions 

are limited, yet they are existent. 

Carltruck cotlisions (945 accidents) 

Carltruck accidents account for over 50go of all 

truck accidents in which people are injured, and 

comprise about 46Vo of all fatalities in accidents in 

which trucks are involved. For the car occupants the 

frequency of fatal injuries in truck,/car accidents is 

roughly four times higher than in car/car accidents. 

Figure 3 gives an overview of the type of collision 

which occurred, at the same time definirrg which areas 

of the car and truck impacted with one another in the 

main collision phase. 

In assessing the ranking order of the types of 

collision, a distinction has to be made between collision 

frequency and the severity of the injuries to the 

car occupants. Car collisions against the sides of the 

trucks (30.490) and truck rear-end accidents with cars 

(26.0V0) are the most frequent. But if the resultant 

injury severity is taken into consideration, it is the 

front,/front collision (collision type A) which is by far 

Typeof collisron {iHr An-\I (ivERALt lNtilrtY }EVtRrrY iN {:îl 

cnr / r uck l :iiiri,ij;!!i.:i ;:;ffii':i;;-"" ll f'rars:3 

I s ll ""û. I 

Front-trontA m]Hflfl- L_ 184 1S.8 QA 

Front-side B 293 30.4 ZDO 

Fronl - reor c tEBl -l:: [D 88 9.4 33 159 

side -fron D 

Recr-{rcnt 

E 

ftrr t3t, al 

t5 4 

ffiffi]- - r 

'rr iltl_]Illf 

.f--l 2t2 260 t, L U 

TotoI 931 r00.0 100 0 

Flgure 3. Type of colllelon ln truck/car accidents. frequency 

of occurrence and serious Inluries 

resulting 

676 


the most important and covers about 40go of the 

cases recorded with serious,/fatal injuries from AIS 3 

and above. Accidents of cars running into the rear 

ends of lorries have a share of about 20go of the 

dangerous./fatal injuries in the accident material. The 

paramount importance of an optimum rear and front 

underrun protection of the truck can be derived from 

this. 

The frequency disrribution of each of the impact 

area$ on the truck is given in fJgure 4 for all of the 

945 car/truck-collisions recorded. By far the most 

frequent are collisions against the truck front (6090), 

which occur namely in the left-hand outside third of 

the truck front. The lorry is struck on the side in 

about one quarter of the cases, 40go of these occur* 

ring in the area between the axles. The impact areas 

on the left are far more frequent than on fhe right. 

Rear end collisions in which the back of the truck is 

struck take place mainly in the left-hand third (55V0). 

The car driver instinctively tries to take evasive action 

to the left, the main load to the rear underrun 

protection thereby being not only concentrated on one 

side but also quite often at an angle as well, which 

results in the problem of bending away the outer rail 

of the underrun protection and thus higher strength of 

the underrun protection device is needed. This risk, 

which is still quite considerable, is reflected by the 

fact that accidents in which cars strike the rear of a 

truck have a frequency of altogether g.5go only, but 

comprise l1vlo of all fatal accidents in the case of 

carltruck accidents. 

There is no doubt that even with the improvement 

of the reinforcing members of the truck rear underrun 

protection by the revision of the ECE-R 5g /Z/ there 

is definitely not sufficient prorection. This would 

require above all that the rear underride protection of 

the truck should be lowered to about 30 cm above the 

ground and the permissible distance from the rear 

edge of the loading area should be reduced. Further 

improvements must-as already mentioned-be made 

to the reinforcing support members in the event of an 

angled impact in the outer area and increased test_ 

loads compared with the actual ECE-R 5g are to be 

required. The relative impact speed of the car on the 

rear of the truck is in most cases in the region of 

about 30 to 50 km/h /i.41. 

$ffiffi,tu 

''"mffi&ffiH 

Flgure4. 

Dlstribution of the impacted areas of the 

truck in colllslons with passenger cars (94S 

accidents)

In the case of the dominant front,/front collisions, 

quite naturally serious difficulties are caused by the 

masses involved and the high relative speed. But 

studies /3,5/ showed that even serious lrontal collisions 

were mostly not accidents with any offset, in 

which, of course, the necessarily high truck mass fully 

loads the opponent's car: heavy car-truck collisions 

are mostly accidents with comparatively large offset, 

in which, however, the front design of the truck with 

high mounted bumpers and rigid structures results in 

an unfavourable application of force against the car 

(Figure 5). The truck overruns the deformation structures 

of the car provided to absorb €nergy, forces the 

car under the truck front and wedges it there (Figure 

6). This fact in an overproportionate number of cases, 

result.s in massive intrusion of the passenger cell with 

very serious injuries being caused to the car occupants. 

Since 1978, the HUK-Verband together with the 

Technical University in Berlin and truck manufacturers 

has continually carried out car crash tests into the 

frolrt of trucks to develop counter measures based on 

the knowledge of real-life accidents /6/. 

These series of tests showed that to achieve the aim 

of truck front underrun protection, both measures 

with regard to the truck l'ront geometry and the front 

deformation characteristics are nece$sary. 

Safety improvements to the truck front design demand 

first of all 

t a lowering of the bumper to a clearance of 

about 30 cm. so that the car structures are 

no longer driven oyer by the truck and the 

car becomes wedged underneath the truck 

a a larger contact area fbr the car on the 

truck. so that a wider area of force load is 

given and a deflection functiott can take 

, , place as far as possible 

Simply avoiding the car structures being overrun in 

itself results in an advantage, but for a truck front 

protection a further prerequisite is that 

the truck front is designed for a certain 

energy absorption and not, as is the case 

today, so that the deformation work has to 

be performed almost exclusively by the car. 

Figure 5. Typical frontal colllsion of a car into a truck 

with relatively high offset but serious intru' 

sion of the cer 


i;":*\r, 

tr:' Illiftd,rtl 

&' 

I 

I,{10 

Flgure 6. Car to truck crash test: typical force load to 

the car with the truck's lront overrunning 

the deformatlon structure of the car 

As a fundamental finding of the experiments to 

date front structure$ of this kind can be createdalthough 

with some expenditure-imposing no limitations 

on the present truck construction and potential 

usage and that considerable safety gains are to be 

expected for the car occupants /7 /. 

An "energy-absorbing 

design of the truck front" 

can be achieved by replacing the conventional bumper 

systems by a deformation structure which is mounted 

on a carrier construction (Figure 7a), this carrier 

system, being, in its turn, attached to the frame and 

being designed to resist deforrnatiou as far as possible. 

The experiments have not yet been completed; in 

the light of these latest developments a front underrun 

protection of this kind is likcly to mean an additional 

length of the truck amounting to about 10 to 15 cm 

(Figure 7blc) and additional weight of some 150 kginsignificant 

compared with the total weight of a 

Figure 7. 

, sr i ,'1' ,!rSat1!F 

Possible constructlon ot a front underride 

protectlon 

a) carrler construction and detormation 

structure 

b) truck front underride protectlon; vlew 

from below 

c) truck front underrideprotection; 

vlew 

lrom the side 

6'77

truck, but of not inconsiderable significance for 

economy calculations. 

A front protection of this kind results not only in 

advantages for the car occupants but can also improve 

the safety of the truck. Even comparatively low 

relative speeds often cause a massive car impact 

against the truck's front wheel damaging its steerirtg 

or even making the lorry unsteerable, so that quite 

often very serious secondary collisions occur as a 

result. This risk can also be greatly reduced by a truck 

front underrun protection in the way suggested' 


M 

ffiffi[ ]Et 

Flgure 9. Truck colllelons wlth two.wheeled vehlclee; 

dlstrlbutlon of the lmFacted areas (390 accl. 

dente) 

Collisions between trucks and two-wheeled 

vehicles (390 accidents) 

In the case of these collisions with trucks of 3.5 t or 

more a distinction was made between the categories 

"motorcycle, 

light motorised two-wheel-vehicles 

(Mofas, Mokick) and bicycles". 

The individual categories of two-wheeled vehicles 

show different accident characteristics /3.4/: in one 

third of the cases involving motorcycles, frontal 

collisions against the side of the truck take place 

(Figure 8). When a truck turns off, motorcyclists are 

apparently frequently overlooked or their speed is 

wrongly estimated. In comparison with motorcycles, 

however, "light motorised two-wheeled vehicles" and 

bicycles are struck far more frequently by the truck 

front, usually in cross traffic. 

The truck's side area is involved in about 5090 of 

the cases in collisions with motorised two-wheel-vehicles- 

(Figure 9); more than one-third of all collisions take 

place in the area between the axles (3,4), these 

accidents accounting for about half of all motorcyclists 

killed, as well as a considerable proportion of 

seriously injured people. 

Quite a number of these involve glancing off 

impacts (for example when overtaking) which, accounting 

for a share of ?0 to 33V0, represent the 

subsequent fall into the space between the front and 

rear axles with resulting run-over by the truck. 

The danger of running over increases at a lower 

speed of the two-wheeled vehicle in keeping with the 

proportion of overtaking manoeuvres. For example of 

186 cyclists in the truck accident material with personal 

injuries /4/, 14 persons were run over. 

Figure 9 shows that in truck collisions with two: 

wheeled-vehicles, the right-hand side area is the critical 

impact area*a side underruu protection on the 

truck would influence around 5090 of the serious- 

/fatal injuries to drivers of two-wheeled vehicles 

and-in particular-would totally avoid the dangerous 

falls between the front and rear axles. Any improve' 

ment of the present situation including refitting trucks 

and trailors with side protection results in advantages; 

an optimum side underrun protection, however, 

should be designed with a flat surface covering the 

whole space of the side especially on the right-hand 

side of the vehicle. Thus the fact is also taken into 

consideration that especially motorcycles collide relatively 

frequently head on with the truck side area and 

that therefore the driver incurs the risk of becoming 

entangled by the truck. 

Collisions of trucks f,nd pedestrians (112 

second highest risk at all. Decisive here is not the 

intensity of the impact itself but the danger of a 

accidents) 

Here, of course, the most serious accidents are 

encountered: only one-third of all pedestrians get off 

with only moderate injuries in truck,/pedestrian collisions 

/3,4/. 

The impact of the pedestrian most frequently takes 

il 12 l5 l6 20 13 

place (4290) with the front of the truck in the right 

"L 

r#F4 

nt.at 

I -ffi# 

-.t - 

?_ 

l0 

17 

5 

E? 

rg 

3? 

l1 

28 

1r0 

3L 

_ol t:n 1l 

2L 61 l0 6 

-.-' 

g9 254 lr @ 

5? 1l ll 

s7 1()2 26. 36 tr 

ll 1L 2 I 

t6E 190 ]m0 l5l rm( 

hand area; in over one third of the cases an impact 

against the right hand side area occurs and protective 

measures must first of all be taken in thege areas 

(Figure l0). 

\ilhat is decisive is a closed surface and avoiding 

protruding edges in the vehicle's front and side 

especially between the truck driver's cabin and the 

Flgure 8. Accident characterlstlcs of collision between 

trucks and two-wheeled vehicles, frequency 

ol the areas lmpacted and serious inlurle$ 

resultlng 

loading area. Also by mounting a truck front protection 

as proposed at a short distance from the ground 

certain safety improvements, for example, avoiding 

being driven over, can be carried out.

'@ffiffiffi] 

M M 

-- €I 

Mre 

Flgure 10. lmpactcd arsas ln colllslons ol trucks wlth 

pedeetrlanr 

(1 12 accldentr) 

About two-thirds (65.990) of truck/pedestrian accidents 

take place irt a collision speed range of up to 30 

km/h /3,4/. Because of this relatively low speed the 

search for safety measures, also with regard to pedestrian 

protection, do not $eem to be completely hopeless. 

The cotlision risk for the pedestrian could also be 

reduced by "active safety"; since the pedestrian is 

frequently struck only by the outer parts of the truck 

/4,8/, equipping the truck with anti-locking brakes 

(ALB) a reduction could be expected because of the 

possibility of steering the truck in spite of an emergency 

braking. 

Safety measure$ for the truck occuprnt$ 

(sample 770 rccidents) 

While in the last few years an increasing number of 

studies has been made on the subject of "partner 

protection" few studies are available today on the 

risks to the truck occupants. For this rea$on new 

uninjured 

Number 


$li8htly 

inlutrd 

Numbef 

accident material from 770 collisions involving trucks 

of 3.5 t or more was evaluated, the studies covering 

the whole of the Federal State of Bavaria, which 

means they can also be regarded as representative of 

the accident situation in Germany. Altogether 3,500 

truck accidents from 1984 were evaluated, and from 

these the above-mentioned accident material II was 

selected according to the following criteria: 

To describe the current state of truck safety, only 

accidents involving a truck of 3.5 t or more, manufactured 

in or after 1976 were included. In addition to 

the injuries to the occupants in the truck accidents 

with safety related damage to the driver's cabin but 

without injury to the occupants were also included in 

order to avoid a negative selection and thus to 

objectively reflect the safety risks to the truck occupants. 

The study showed that the injury risk of the truck 

occupants is almost exclusively dominated by truck/ 

truck collisions and single-truck-accidents (Figure ll). 

Carltruck collisions are, of course, frequent as far as 

the number of cases is concerned and can also result 

in considerable front damage to the truck, but the 

resulting injury risk of the truck occupants is nevertheless 

slight and the serious injury consequences are 

mainly due to secondary collisions. 

Surprisingly not the lighter mass categories of up to 

12 t were found to be overrepresented in accidents 

with injury to truck occupants in the accident material, 

but mainly the heavy trucks of 16 t or more, the 

truck trailers or the articulated lorries as compared 

with their proportion of the registrations. Thus, the 

priously 

Iniur€d 

Nrrmbcr | 

% 

Fatalities 

Number % 

TO1'AL 

trurk trcupdniS 

Number | % 

SinBle truck ffcid€art 19 115 L3 33.6 L 30.8 1B'l 18.8 

fruck/truck 80 136 57 44.5 6 46.1 279 291 

Truck/car 325 1A7 20 r5.6 3 23.1 4s5 l'7.3 

Truk/othr:r 26 12 I 6.3 /.6 1.8 

TOTAL 450 370 128 100.0 13 100.0 961 

100 0 

Figure 11. Truck accidents with injury or serious $aiety riek to truck occupants, resultent risk 

related to type ol accident 

679

"weight 

bonus" of heavy trucks involved in collisions 

with other accident opponents seems to represent an 

additional burden in truck,/truck collisions and single 

truck accidents /9/. 

Two out of three truck/truck collisions with injury 

to the occupants are frontal collisions with the rearend 

of another truck, which in 56o/o of the cases 

occurred on motorways, and most of which involved 

long distance trucks (Figure l2). In 72.10/o of these 

front/rear-end collisions the relative speed was under 

30 km/h, and in 9690 under 50 km/h. 

Figure l3 shows the darnage characteristics in truck/ 

truck collisions in general. The intrusion in rear-end 

impacts of trucks*being the most predominant risk 

for occupants-is shown in Figure 14. Even if the 

driver often tries to nrove over towards the left as f'ar 

as possible at the last moment, the outside left-hand 

third of the front is not less frequently loaded with 

direct deformation. The central problem for the 

stability of the driver's cabin and thus for the injury 

risk to the occupants is connected with an impact 

against the protruding, rigid and not compatible 

rear-end of another truck /4/. 

The dangerous intrusions of 60 cm and more were 

not to be found in cases with partial force load, but 

in full force load to the whole cabin without offset 

(Figure l4). The main deformation here is not in the 

roof, but in the lower area between the bumper and 

lower frame of the windshield. Even if this lrequently 

results in massive intrusion, normally a ,.minimal 

survival space" remains for the truck occupants, even 

in extremely serious accidents. 

680 

Iype of 

co It ision 

A B 

+ + 

A q obtrrvcd trwk 


MAIS 2-5 

Number I 

gt 

B 4 opponent 

The second greatest risk of injuries to the truck 

occupants is encountered in single vehicle accidents, 

5090 of which are characterized by the vehicle leaving 

the road and then driving into a ditch or running into 

the embankment without a massive collision with an 

object. Only in about 2590 of the cases was there a 

massive collision with a tree, a wall or some other 

object. ln 7lslo of these cases, however, the speed was 

over 45 km/h and often in the critical range. 

The damage characteristics in single vehicle acci_ 

dents compared with trucky'truck collisions show 

clearly that the impact is in the right-hand area in 

4090 of the cases, the intrusion of the driver's cabin 

being generally less oflten severe (Figure 15) in com_ 

parison with front,/rear end truck collisions. 

Special risks are involved when the truck tips or 

turns over, and this alone accounted for 50Vo of the 

fatalities in the rruck and about 25Vo of the injuries 

ranging from AIS 3 to AIS 5; these cases occurred 

mainly in single vehicle accidents. 

In 147 of the 770 cases (Iggo) of this sample the 

truck turned over with the vehicle tipping over in 134 

cases (usually turning over onto the right side) and 

rolling over completely in only 14 cases. When it tips 

over to one ride the intrusion into the driver's cabin is 

usually moderate, but the danger of ejection of the 

unrestrained occupant or the risk of intcrior impact 

by being thrown to the side is very great. 

In this connection no major problems were ob_ 

served with rhe rigidity ot the roof-although there 

was considerable local intrusion in some cases. the 

minimum survival space was maintained even then. 

MAIS 6 

N{rmber I 

all jniured 

TOIA{ 

lauck trcuFantS 

Numbe. I X 

. L|_ 

l-ront/tront -'_1LJLJffi 3 7.0 I (16.7) ?t, I t.3 

Frontl 

side 7 163 2 (33 

3) 27 1?,7 

Fronl/rear 32 7lr./- 3 ( 50.0) 132 6l.s 

Other types of cottisions I 23 30 11,.1 

Toto t 43 r00.0 6 100.0 2'r3 r00 0 

Figure 12. Hiqhest degree ol iniury severity of truck occupants involved in truck.ro-truck 

eollisions-related to type of impact

lftation.)t llx' 

deltrrndtr0il No. ol cit*s frequenr y ol defsmarion 

upp€r area trly 

lower area only 

totnl frontal areJ 

5 

159 

187 


A G caies with partial olfstt 

Figure 16 gives the total injury severity of the truck 

occupant$ involved and the frequency/severity of the 

injuries to the various parts of the body. 

As many as half of all severe/fatal injuries to the 

truck occupants in this study are due to ejection. In 

890 of tlte 710 cases intrusion was so great that the 

occupants were wedged in the vehicle (usually in the 

region of the legs). The other injuries occurred mainly 

through impact in the interior, the "steering wheel/ 

steering column" and dclormation of the "foot/ 

leg-space" dorninating as the cause of serious injury. 

The impact to the windshield is frequent but mo$tly 

without scrious consequences. Occupant impact to the 

truck's side areas and in the roof account for a 

comparatively Iow proportion /9/. 

The injuries to the individual parts of the body 

clearly confirm the injury risks typical of trucks. 

Head injuries are bound to dominate in fatal injuries, 

but abdominal and chest injuries are almost of equal 

significance. Among the serious/critical injuries of 

AIS 3-5 injuries in the abdominal, chest and leg areas 

occur far more frequently than head injuries-an 

indication of the problems connected with occupants 

being wedged in and the injury causes connected with 

the steering wheel,/steering column. In the case of 

B = ca*s without oft*t 

Flgure 13. Frequency of damaged arear of the truck'g front In truck to-truck collislone generally 

intrusion the unfavourable impact conditions of the 

steering wheel-flat normal position-thus locally 

concentrated impact loading and only limited yielding- 

are, in addition, further aggravated by tlre fact 

that the steering whccl,/steering column are frequently 

pushed upwards (Figure l7)- 

Assessments were made to find out which safety 

measures are suitable for reducing the comparatively 

low, but existing injury risks ol' the truck occupants. 

The best protective effect could be expected fronr a 

safety belt, which in approximately 50a/o of the cases 

would result in a considerable reduction of injuries. In 

particular the fatal injuries (ejection) and the serious 

injuries in the abdominal and chest region as well as 

the thigh fractures as a result of the occupants being 

thrown forward would thus be considerably reduced. 

Problems which were at first expected, for example 

caused by excessive intrusion and lack of survival 

space, were-except in two ca$es*not observed in this 

study and are below I go. But it must be pointed out 

when discussing the question of safety belts that belt 

characteristics (restraint effect as with a 3-point-safety 

belt) and function (taking into account the truck's 

vibration) rnust not restrict the ergonometry.

E 

z 

; 

D 

= 

Totol 


Domoged oreo 

no isolated local intrusion 

only in combination with lowcr area 

Flgure 14. Range ol lntruelon to th6 dlffer€nt areas of the truck lront In truck-to-truck rear end 

collisions 

location oI thc damage 

upyrr area only , ZO 

lower arca only 

total frontal area 

WZ/ffi 

E 

z 

= 

20-40 

40-60 

'60 

-20 

20-40 

.r0 - 60 

560 

_ ZO 

20-+o 

,to - 60 

>60 

t-)anra6ed area considcrcd 

Do mo ged oreo 

front third only . combined area .no offset 

Totqt 

1 r z r s l + t E l e 

2 2 

2 ? lr 

1 2 3 

1 I 1 1 1 7 11 

I 3 t, 

6 6 

1 /, 5 

Totol lr 1 1 1 t, 2t, 35 

Flgure 15. Range of intrusion to the dilferent areas ol the truck front in single vehicle accldents

PARTS tr THE 

DODY 'NJUNED 

Nek/cffvErl ry# 

IPt_r.B__Ejll3gqIIE5 

- Chrl 

- B.cl d 

L(:4!8- EX TR EMrTrEs 

- Thith 

. Lfls rrt 

Totol AtS 1- 2 Ats 3- 5 AIS 6 

I g/. 

- * l 

t{ 

6r 

26 

, 1 

nl 

31, 

13 

36 

75 

t9 

38 

379 

76 

64 

Il9 

5.1 

9? 

lzl 

66 

8,. 

70 

I 1,6 

7L 

r83 

36 

33 

1,5 

25 

62 

33 

2L 

6S 

69 

37 

39 6 

IE 

7.1 

9.7 

101 

I J.t 

7.1 

63 

5? 

119 

11,$ 

0.0 

32 t2 32 69 

Efari4 90 r76 s0 19,5 

Toral: injurd 

Flgure 16. Frequency and EcyGrlty ol Infurlu to tho 

dlfferent perts of the body ot truck occupants 

Active $[fety-avolding truck nccidents 

(sample 300 accidents) 

A representative sample of 300 accidents were 

selected from accident material II in which safety risks 

for truck drivers, including car/truck accidents were 

analysed. Collisions with two-wheeled vehicles/pedestrians 

were deliberately not irtcluded here, since these 

Flgure 17. Exrmples of typical intrusion of the, drlver's 

cabin, llluetrating the injury risk from 

the steerlng assembly and in the leg ar€B 

l_ 

2 

t2 

I 

; 

I 

10 

6 

t0 

I 


10, I 

RI 

^:, 

J l,tt 

21 

21 6 

270 

162 

270 

tt 

7 53 I 

| 77 

l. 308 

E 

I 

t6l 

5l? r00 0 162 r00 0 37 100.0 l3 r00 0 

5r2 

(19 

sccidents rcveal completely different characteristics 

with different safety requirements (for example, visibility) 

and will have to be treated in a separate study. 

The criteria for inclusion in the present study 

material are given once more: 

I trucks of over 3.5 t maximum laden weight 

manufactured in or after 1976 

t injury to the truck occupants and/or damage 

to the truck driver's cabin which is of 

significance for safety. 

The high proportion of truck-trailer combinations 

and articulated trucks, altogether 55.0V0, (figure 18) is 

striking in this accident material. Even if the selection 

criteria had a certain influence, this does show the 

great significance of truck-trailer combinations and 

articulated trucks with respect to active safety and this 

confirms that in the case of these truck categories all 

possible technical efforts will have to be made to 

reduce the accident risk. 

An accurate assessment of questions of active safety 

presupposes detailed knowledge of the individual precrash 

situations and the typical behaviour patterndata 

which has up to now only been available to a 

limited extent and only rarely for typical truck categories. 

Experience to date shows that truck accidents can 

be classified iflto a few main types which recur with 

similar crisis situations. 

With regard to the recording criterion the following 

frequency of dominating "crisis situations" was to be 

found (Figure l9): 

I rear-end accidents on Autobahns (24.390) 

. single truck accidents (20.3V0) 

I accidents in the area of crossings (21.390) 

Truck typ+ 

(rotal ladcn s?itht) Numbcr $ 

Smnll truck :7r5 t 97 2t,9 

limpfq 1,t.k 7., - 12 | 15 39 

simDle truck 'l? t 63 l6 2 

Tru.k wrth irailcr ;28 t 31 80 

Ir(:k with trailer>?t : ft t 95 21.4 

Arliculated lorries 8B t L.o 

fo(nl of truckr rf,valved 389 r00 0 

5s0% 

Flgure 18. Representative sample of 300 truck accidents 

( = 3,5 tonnes); selection crlteria: 

cases with injury to truck occupants or 

safety-related demage to the driver'e cabln 

, mass category of the 397 trucks lnvolved 

683

m-coming traffic 

accident, straight 

orirorning traffic 

accidenl bend 

Collision at 

crossing/jrrnction 


Totah 300 at:t:idents 

Figure 19. Frequency of critical pre-crash situatione 

It is typical of the pre-crash situations termed 

"oncoming traffic accidents straight" (l4.7Vo) and 

"oncoming 

traffic accidents bend" (10.790) that most 

of them are not characterised by wrong behaviour on 

the part of the truck. In three out of four cases the 

accident opponent-mostly a car- crossed onto the 

truck's side of the road through skidding, overtaking 

or excessive speed, unfavourable road conditions, 

wetness or snow, being present in about 50 to 7590 of 

the cases. 

Further details of the "crisis situation" result from 

the breakdown by vehicle categories (Figure 20). By 

far the largest proportion of the truck-trailer combinations 

and articulated trucks in the accident material 

r.'. I' i i / | /////fi 7/, 7I i'////r, 

: t 

// / r ;////// r f f/ f.///;/f /. 

,',' )'. / i,m 

/t l/////r't'////'/./ /, // 

I 

// 

Rcar+nd (:ollisiorrs 

Autobahn 

Rcar-cnd

ear-end accldents caused by false estimates of slow 

trucks could be avoided. 

The dominant occurrence of rear-end accidents on 

Autobahns also led to an analysis of the possible 

effect of a "distance warning device" for trucks. As 

was expected, it was shown that they would be of 

Breat benefit: 3l9o of the truck accidents on the 

Autobahn would certainly have been avoidable and a 

further 220/o might have been avoided or at least the 

resulting damage could have been reduced' On coun' 

try roads this figure is reduced, although it is still as 

high as l2.4Vo. 

It would be a pre-requisite, however, that the truck 

driver observes this information. The problems with 

overtaking, cutting by other vehicles, etc., are obvious- 

but if the development of a feasible distance 

warning device for heavy trucks, could be successful, 

a considerable proportion of the most serious accidents 

today could be avoided. 

In further studies the braking behaviour depending 

on the crisis situation and the possible benefit of an 

antilocking brake system (ALB) were examined 

(Figure 2l). The study showed that in 52'690 of the 

ca$e$ an emergency braking is made by the truck 

driver and that-especially in accidents at crossingsthere 

is, even today, an attempt to react by braking 

and steering. But as the wheels are locked this is not 

only unsuccessful but also involves the danger of 

secondary collisions. 

In a detailed case study the accidents were evaluated 

with regard to the possible benefit of an anti-locking 

brake system (ALB) using the definition: it was 

assumed that the accident would certainly have been 

avoided if the driver tried to react by steering and 

braking which would have resulted in avoiding the 

accident but which was not possible because the 

wheels locked. It was assumed that the accident would 

probably have been avoided as the accident situation 

had providecl a clear possibility of doing so if the 

driver had used a steering reaction in spite of the 

emergency braking. The effect of reducing the accident 

consequences was only taken into this category if 

tlr{+d (alliiimi, Auld$n 36 

Eo *mrtfry'y Frtsrnt 

i''", ;"; ;i*' "', 

l' 


v,ilr emcrf,cEy F*trd 

rorar 

I ol rrrq llcr.'{ 

| "., 

H.n...rd .olLr'dE 

o I 5 20 

18 

M..omint lralln .r.i&ilt 

-itIE;s':T -- l_n 

2L 

I 

27 3fi 

l1 7 1E I 

\tr'Rk truk &fffit t1 12 rS J 

I 5 

1l 

71 

* 

! 2 

t, G 

'il dll i rrsrs ersims 137 37 100 152 L3 r09 

r0 

t3 

'11 

t2.tt ll.6l, u.0 t 37.rt 

Flgure 21. Emergency reaction of truck drivers in the 

critlcal pre-crash phase 

a v€ry substantial reduction of accident intensity could 

have been expected with an anti-locking brake system, 

It was shown that high effectivity would exist for 

Autobahn accidents of trucks by using an anti-locking 

braking system about 7Vo of the truck accidents 

would certainly have been avoided and a furthet l9Vo 

probably. This high rate of avoidability is above all 

influenced by the proportion of accidents involving 

truck-trailer combinations, in which the trailer fre' 

quently gets out ol control if an emergency braking is 

made. 

ln the case of truck accidents on country roads the 

figures were lower because of the different traffic 

characteristics, but even so 3.lVo of these accidents 

could certainly and an additional 10.890 could probably 

have been avoided. If it is considered that irt 

about half of the accidents the crisis situation occurs 

so quickly and unexpectedly that an effective braking 

reaction cannot take place the anti-locking brake 

system offers a high degree of effectivetress. 

These findings confirnr independent studies into 

accidents of trucks with dangerous loads /10/ and bus 

collisions /ll/ in which it is estimated that the use of 

an anti-locking brake systcm would certainly avoid 

590 and probably l59o of the accidents. 

From a technical point of view fitting heavy trucks 

(12 t and up) with anti-locking brakes as standard 

equipment-especially, of course, articulated vehicles 

and truck-trailer combinations-is the principle mea' 

sure for reducing the existing braking problerns. With 

regard to the accident risk, however, in a second 

phase the smaller trucks, too, should be fitted with 

anti-locking brakes as standard equipment in order to 

avoid or reduce accidents especially in the crisis 

situations in crossings-frequently accompanied by 

extremely serious itrjuries to the accident opponents. 

The intention of the West German Covernment to 

make anti-locking brake systems compulsory standard 

equipment for trucks in the future is a decisive step 

towards more safety for trucks. 

In further studies based on the present accident 

material the peripheral conditions for the demands 

made on the anti-locking brake systems will be 

analysed. A warning, however, must be made against 

exaggerated hopes, since au accurate recording of the 

different parameters, such a$ Ioading conditions' 

changes in friction values in a longitudinal or transverse 

direction is very difficult and is often even 

impossible. Analysis of braking marks and the precrash 

frequency of trucks leaving the road and driving 

on to the shoulder lead us to expect, however, that in 

l09o of the accidents p-Split-conditions probably 

existed. For this reason, if trucks are to be fitted with 

anti-locking brakes as $talldard equipment, an optimum 

system of Category I in accordance with the 

ECE regulation R 58 /2/ should be aimed at. 

685

Summary 

685 


Truck accidents represent a significant factor in the 

overall accident scene: every sixth fatal accident involves 

a truck; each year in Cermany about 1,700 

people are fatally injured in truck accidents and 

40,000 are slightly/seriously injured. 

With regard to passive safety (partner protection) 

the measures termed t*truck front underrun protection" 

in carltruck accidents and mounting a ,,side 

undoubtedly represents an improvement, although 

there are still problems in the case of offset impacts at 

an angle Onto the outer areas with the risk of the 

outer part being bent off. An improvement in the 

supporting reinforcing members and increased test 

loads in ECE-R 58 (150 KN instead of 100 KN) would 

be desirable; the rear underrun protection should have 

a clearance from the ground of no more than 30 cm 

and be mounted as near as possible to the rear end. 

underrun protection" in collisions with two-wheeled 

vehicles and pedestrians are of paramount importance, 

but the safety criteria for truck occupants 

should not be ignored either. 

With regard to the passiv e safety of truck occupants 

(self-protection) measures are possible and necessary 

in respect of a truck structure, the design of the 

interior and the use of restraint systems. 

In car/ffuck collisions the problem zones are to be 

found in the area of the truck front, in the outer 

left-hand third of the truck's rear end and on the 

truck side between the axles. 

About 6090 of the serious carltruck collisions 

The main problems connected with intrusion do not 

occur in truck collisions with objects but in collisions 

with other trucks. The decisive intrusion zone is 

therefore in the region between the bumper and the 

lower windscreen frame-typical of the collision with 

involve the truck front. By creating a truck front 

underrun protection system instead of the rigid 

bumpers, usually mounted at a high level nowadays, a 

considerable reduction of the risks to the car occupants 

can be expected. 

The development work to date shows that-although 

with some expenditure (costs, 150 kg weight, 

about l0 to 15 cm additional length)-effective solutions 

are possible by mounting energy-absorbing struc- 

the rear of a lorry. With regard to roof rigidity the 

present study does not reveal essential safety defects; 

in spite of, in ssme cases, massive intrusion, a 

necessary minimal survival space remained, 

Safety tests of the structural rigidity of the driver's 

cabin should therefore be carried out against structures 

similar to the rear structures of trucks /lZ/, 

Isolated loading of the roof does not correspond to 

real-life risk situations. Turning oyer onto the side or 

tural elements onto the carrier structure of a truck 

front pfotection as proposed. 

A front protection of this kind would have the 

advantage for the truck itself that in the event of 

accidents of considerable severity no direct impact 

would occur against the front wheel and this would 

avoid problems of damaging the steering and,/or the 

steerability (danger of secondary collisions). 

By being mounted at a low level, the front protection 

could act positively by increasing a deflection 

effect in collisions with two-wheeled vehicles and 

pedestrians who are quite often today knocked over 

and driven over; moreover some elastic absorption 

area could be attached to this delormation $tructure, 

and this might cushion at least to a limited degree the 

main points of impact. 

The principal requirement for passive safety in 

collisions with fwo-wfteeled vehicles and pedestrians, 

however, is to mount a side underrun protection and 

to consistently make the truck less aggressive in the 

region of the right-hand front and side area. The side 

underrun protection should be designcd with a plain 

surface in the total area between the front and rear 

axles, particularly since it should be effective not only 

in the case of glance-off collisions, but also in the 

case of relatively frequent impacts at an angle by 

motorcyclists in this area whcn the truck turns off. 

The rear underrun protection device in its present 

design in accordance with ECE regulation R 58 

even complete rollover, only rarely result in massive 

deformation of the driver's cabin and serious restrictions 

of the interior-the central problem is, rather, in 

the intrusion of the front area between the bumper 

and lower frame of the windscreen. 

In the truck's interior, energy-absorbing padding 

should be improved as it provides better protection 

against leg injuries to the driver, and avoids, as much 

as possible, his being wedged in in the event of 

intrusion of the front structure of the driver's cabin. 

The flat position of steering wheels nowadays 

represents a considerable injury risk to the chest,/abdominal 

region for the unrestrained truck driver. It 

should be proved whether truck steering wheels can be 

better designed from the biomechanical point of 

view- at least design measures must be taken to 

avoid the dangerous pushing upwards of the steering 

assembly in the event of intrusion of the fiont 

structure. 

The risks for truck drivers which dominate today as 

a result of ejection and impact in the interior, 

especially aI the steering column,/steering wheel, 

would be considerably reduced by wearing a safety 

belt; with safety belts-adapted, of course, to the 

specific requirement of trucks-a probable benefit 

could be expected in about 50go of the accidents 

recorded here, reducing in particular serious/fatal 

injuries to a greater extent. A negative effect, if the 

occupants are restrained in their seats, in the event of

intrusion did not prove to be relevant in this study 

and seems not to be different to the existins situation 

in passenger cars. 

In the field of active safety the standard equipment 

of trucks, especially trucks of 12 tonnes and up, 

truck-trailer combinations and articulated lorries with 

anti-locking brakes constitutes the central safety measure 

of the next year$. It is expected that the 

possibility of being able to steer and brake in emergency 

situations would certainly avoid 590 and probably 

l59o of the truck accidents and considerably 

reduce their consequences,-especially since these accidents 

frequently occur at relatively low collision 

speeds, but nevertheless result in extreme damage by 

reason of the great masses involved. But in a second, 

later equipment phase the lighter lorries (starting from 

3.5 tonnes and up), too, should also be equipped with 

anti-locking brakes, since especially lighter trucks are 

involved in accidents with considerable injury to 

people as a result of their being used in built-up areas. 

The great significance of accidents on Autobahns 

for heavy trucks was confirmed, rear-end collisions 

dominating, As an additional preventive measure to 

avoid accidents, an improved rear signalling system 

for trucks (warning for vehicles behind when trucks 

are driving unusually slowly on Autobahns) should 

also be examined as well as the use of distancewarning 

devices, especially on Autobahns. There are 

certainly considerable problems in the practical operation 

with these two possible mea$ures; by reason of 

the serious risks in these frequent critical situations, 

however. solutions must be achieved. 

The catalogue of measures described for active and 

passive safety must, of course, remain incomplete. 

But it has been possible to acquire a large body of 

representative material on truck accidents, and in 

further studies possible individual measures will be 

examined. The practical applications of the measures 

mentioned woultl, however, already result in a considerable 

improvement in active and passive safety in 

accidents involving trucks. In this connection intcrnationally 

uniform safety regulations should be available 

for trucks in order to achieve the same safety standard 

and uniform conditions of competition in the 

international traffic of today which will increase in 

future. But these regulations nust guarantee a high 

level of safety and must not represent a minimum 

common denominator. 

Even with an increase in the transport requirements 

of the truck, €conomy and safety are not contradictory 

demands. Studies of truck accidents must be 

continued and intensified to present criteria for deciding 

on future technical developments and possibilities 

of further advanced truck safety. 


Concluding Remarks 

The authors are indebted to all the member insurance 

companies of the HuK-Verband in particular for 

making their claims records readily available for this 

study, and to Messrs. Heider and Zistler, members of 

the HUK staff who carried out accident analysis and 

evaluation work. The authors also owe a debt of 

gratitude to the various police departments. 

References 

l. Statistisches Bundesamt Wiesbaden, Reihe 3.3, 

1984 / SS " Stradenverkehrsunfdlle" 

ECE-R 58; "Einheitliche Vorschriften ftir die 

Cenehmigung von Nutzfahrzeugen, Anhflngern 

und Sattelanhdngern hinsichtlich ihres rtickwdrtigen 

Unterfahrschutzes", Economic Commission 

of Europe 

K. Langwieder; F. GauF; W.D. Schmidt; M. 

Wrobel: "AuBere Sicherheit von Lkw und Anhdngern", 

Forschungsvereinigung Automobiltechnik. 

Schriftetrreihe Nr. 27 

K. Langwieder; M. Danner; M. Wrobel; "A 

Contribution to Risk Analysis and the Characteristics 

of Truck Accidents", XX. FlSlTA, Congress, 

Wien, May 1984 

M. Darrner; K. Langwieder: "Results of an 

Analysis of Truck Accidents and Possibilities of 

Reducing Their Consequences Discussed on the 

Basis of Car-to-Truck Crash Tests", Twentyfifth 

Stapp Car Crash Conference, San Francisco, 

September l98l 

a) l0 Lkw Crash Tests; b) Second Seria of 

carltruck crash tests 1984-87, (not yet published), 

HUK-Verband, Munich 

M. Danner; K. Langwieder: "Lkw-Frontschutz- 

ein wescnt licher Beitrag zu mehr Partncrschutz", 

TUV-folloquium 

"Nutzfahrzeug 

2000", December 1986, Koln 

K- Langwieder; M. Danner; W. Wachter; Th. 

Hummel: "Patterns of Multi-Traumatisation in 

Pedestrian Accidents in Relation to Injury Combirrations 

and Car Shape", 8. ESV-Conference, 

Wolfsburg, October 1980 

Research project: "Passive Safety of the Driver's 

Cabin of Trucks", Report to be published in 

cooperation with FAT, Frankfurt, 1987/88 (yet 

unpublished) 

10. K. Langwieder: "Unfhllauswertung bei Gefahrguttransporten", 

DEKRA Fachtagung, Gefahrguttransport 

auf der Starape, Wart, October 

r986 

ll. K. Langwieder; M. Danner; Th. Hummel: "Col- 

2. 

3. 

4. 

5. 

6. 

7. 

E. 

9. 

lision Types and Characteristics of Bus Accidents-Their 

Consequences for the Bus Passengers 

and the Accident Opponent", Tenth 

ESV-Conference, Oxford, July 1985 

687

12. ECE-R 29r "Einheitliche Vorschriften fiir die 

Genehmigung der Fahrzeuge hinsichtlich des 

Schutzes der Insassen des Fahrerhauses von 


Nutzfahrzeugen", Economic Commission of Europe 

Typology of Traffic Accidents Concerning Cars Impacted by Trucks 

Giles Vallet, 

Michelle Ramet, 

Dominique Cesari, 

Claude Dolivet, 

Inrets-LCB, 

France 

Abstract 

The road traffic mixes in the same flow*light cars 

and heavy trucks: well thcn a large part of severe 

traffic accidents concerns trucks impacting light vehicles. 

Actually, if this type of accident does not occur 

freqrrently (4.5 percent of all accident$), it is very 

serious and 10 percent of traffic deaths are due to it. 

Using the sample o[ our bidisciplinary accident investigation, 

this study, concerning 53 cases of actual 

accidents, tries to precise the typology of truck-to-car 

accidents ancl thc influence of the type of irnpact on 

car occupants lesions. Then it determines the more 

representative impacts conditions. 


This study is based on 53 cases of actual accidents 

in which a passenger car was impacted by a truck. 

Flgure 1. Welght of trucks 

688 

l0 

c 

l 

7 

0 

Number o 

I 

I 

8 

I 

0 

These data were collected over a period of five 

years from 1982 to 1986, by the bidisciplinary accident 

investigation team of the L.C.B.(Inrets France). 

We can explain the limited number of cases by the 

fact that, on the one hand, the bidisciplinary study 

collects all types of traffic accidents and, on the other 

hand, it is often impossible to trace the truck, which 

leaves the scene of the accident quickly if it is only 

slightly damaged. 

The aim of this study is to establish the origins of 

crash severity, for this type of traffic accident. 

We consider as truck, any heavy good vehicle of 

which the full weight is higher than 3,500 kg and less 

than 38,000 kg. Among these 53 heavy trucks, we 

have only four coachs, the 49 others are 27 articulated 

and 22 non-articulated trucks. In the 53 passenger 

cars, we have 86 passengers involved, 53 drivers and 

33 passengers. 

Some Considerations Coneerning 

Characteristics of Trucks Involved 

Truck Mass flnd Pessenger Cflrs Mass 

Figure I introduces the actual truck mass at the 

accident time, i-e, the truck weight plus load weight. 

t a q i l r t t t g l n l l H t n t 

Weight (t)


We can see in our sample an approxirnately regular 

distribution between 3,500 kg to 12,000 kg, concerning 

one-third of this sample. Then, we notice a peak 

of 12,000 kg and 16,000 kg truck weight, rhen a 

uniform distribution between 18,000 kg and 26,000 

truck weight. Finally, we notice the highest peak 

(38,000 kg) usually fbr rractor semi-trailers. 

We observe that we have no trucks between 26,000 

range Of mef,,gurements is very extensive between 38 

cm and 70 cm. 

For the passenger cars, the distance from bumper 

top to ground is around 50-52 cm for recent vehicles. 

Impact Speed 

Clearly, this is one of the most difficult frarameters 

to determine, and we may not therefore generalize on 

kg and 37,000 kg. 

impact speeds for each vehicle. However, we were 

The articulated trucks (tractor and semi-trailer) 

(38,000 kg) represent about 20 percent of our sample. 

able to establish in 13 cases, the truck speed immediately 

before impact. This measurement came from the 

The absence ol' trucks between 26,000 kg and chronotachygraph logdiscs. Unfortunately, these log- 

87,000 kg can al.so be explained since this type of discs are difficult to read and since the time between 

truck is used essentially on building sites. 

the beginning of braking and the impact is usually 

When we consider the weight plus load of passarger 

cars (fig. 2), we notice that an important part of them 

extremely short, the speed registered on the disc is 

often higher than the actual speed on impact. Morc- 

are medium sized, indeed even small sized. 

over, the logdisc is often missirrg or illegible for 

This distributiorr is, in fact, very repre$entative of various reason$. 

French number of cars on the roads, 

We have thus 13 speeds varying between 50 and 85 

Distance from Ground to Bumper for Heavy 

Trucks 

This parameter is very important because it deter, 

mines the underruning of passengers cars under the 

truck. 

In fact, one of the most important factors is that, 

when a truck collides with a passenger car, it generally 

km/h, and one case ol'30 km/h. The average speed is 

63,8 km/h with a standard deviation of around 15,3, 

i-e. 67 percent of the speeds recorded are between 48 

and 79 km/h. These speeds (even allowing for them 

being slightly higher than reality) are indeed high, and 

especially if we consider that for those cases where the 

log disc was available, the striking passenger cars were 

still moving. 

hits the car above the bumper i-e, the deformable part 

of the front end, or even the engine part. 

We have measured the distance ground,/truck 

bumper for our truck sample. These mea.surements 

We will have the opportunity to return ro the 

problem of impact speed. 

Different Impact Types 

were carried out on fully loaded trucks. We established 

two categories: articulated and non-articulated 

trucks as can be seen on fig. 3 and 4. 

We categorized accidents according to impact type 

for each vehicle; frontal, rear on lateral, and established 

5 classifications (t'igs. 5 and 6). Bv Fronto- 

The averages are very similar and are situated Lateral, for example, we mean frontal shock for the 

around 54.5 cms. On the other hand, we note that the passenger cars, and lateral for the HCV, The class 

7 

0 

E 

Number 

c 

Flgure 2. Weight of passenger vehicles 

I 

I 

0 

0.6 0.66 0,t o.al 0.? o.tt 0,a o,& 0, 0,a6 r ld||[ r,l I,16 rJ r,$ r,t r.t6 r.a r.a6 r.! 

Weisht (t) 

689

Height (cm) 

Figure 4. Bumper height of 13 tractor-treller unlts 

Number 

llfrE ril 

Flgure 5. TrucUpassenger vehicle lmpact type$ 

690 

Hetght (cm) 

25 

?0 

t F 

l0 

5 

0 


Number 

Flgure 3. Bumper helght of 33 non-artlculated trucks 

Number 

- BIJIPEF ITIGHT tr 33 

HON AFTICIJLITED 

IHUCXS 

--. AVEN^GE 8IJI#EF IIEIEI{I 

_ BUI4PEF HEIGHT OF 13 

TNACTOF-TFAILEA UNITS 

... AVEFAGE EUI{FEH HEIGHT

Number 

t 1, 32X 

15,09X 


g, 431 

22, 64I 

Flgure 6. Trucks passsnger vehicle lmpact typee (96) 

entitled "vs1isqs" includes impacts difficult to classify 

and "rear-frontal" shocks, 

The most frequent impact type is the Fronto- 

Frontal type (42 percent of all cases), which most 

frequently takes place after a loss of control by one of 

the drivers, or after overtaking. 

Fronto-Lateral and Latero-Frontal types have been 

separated in our classification, $ince the lesion mechanisms 

for car occupants are completely different. 

Howsysl, these two types of shocks do have in 

common the fact that they most frequently occur at 

road intersections. 

They represent respectively 27.6 and 15.l percent of 

all cases, i-e, when combined, almost 38 percent of all 

cases, followed by Fronto-Rear impacts (iust over l0 

percent) and the "various" category )also around l0 

percent). Each impact type will be studied separately. 

Impact Direction 

By '*Impact Direction" we mean the direction of 

the main force applied to the vchicle during impact. 

This direction is indicated on a horizontal plane, by a 

clock face reference system "12 o'clock" representing 

the direction along the axis of the vehicle from the 

exterior to the vehicle (Diae. 1). 

,n- 

tn\ 

ti 

/on ./in 

t2h + th 

,'rn/ ,o! ,l ). 

Diagram 1 

\rn 

41,511 

fi 

N 

E 

! 

I 

Fron tofrontal 

Fnon t olBterd 

I 

Latef0f 

nonta 

I 

Fnonto-rean 

Van ious 

However, it must be noted that this orientation of 

force does not allow us to anticipate impact area: it is 

quite possible to have direction " l l o'clock" and 

impact on the left-hand rear wheel. 

Figs. 7 and 8 show the distribution of impact 

directions for different vehicle categories. In both 

cases, the directions ll and 12 o'clock are in the 

majority, which confirms the importance of Fronto- 

Frontal impacts: the most frequent source of these 

directions. Furthermore, for trucks, we observe a high 

number of 3 and 6 o'clock directions, which may 

often indicate lateral and rear impacts. 

Impact Location 

This is shown on figs. 9 and 10, once again with a 

clear majority for frontal impacts. However, we 

observe a difference between trucks and passenger 

vehicles as far as other impact locations are concerned. 

For the passenger vehicle, there is a high 

difference between left and right-hand sides, and few 

impacts at rear, whereas for trucks, neither left nor 

right-hand sides predominate, and rear impacts are 

frequent. 

These different accident characteristics seem to 

indicate the importance of Fronto-Frontat impacts 

truck-passenger vehicle in accidentology. 

The Fronto-Frontal Impact 

This type of crash involves 22 trucks and 22 

passengers car (with 26 occupants). 

Firstly, we exarnined underrun on impact, and 

established four types of underrun for passenger 

vehicles. In the case of underrun, we studied the area 

of the passenger vehicle which had actually been in 

collision with the truck. Our four categories are; all 

the front of the car, two-thirds, one-third, side-swipe. 

Side-swipe produces low underrun, since the two 

69r

Flgure 7. Paseenger vehicle impact directlon 

vehicles tend to move along at a tangent. We refer to 

figs. ll and 12, where for approximately 1/3 of all 

cases, underrun for the passenger vehicle was total, 

and similarly for the category "two-thirds", whereas 

side-swipe represents only 14 percent of all cases. This 

may be explainecl by the fact that these accidents most 

frequently originate from a loss of control by the 

truck or passenger vehicle drivers: and in all the cases, 

we examined in the "all-front" and "two-thirds" 

underrun categories, there were never traces of the 

driver trying to avoid collision, except for some cases 

of last-minute braking. This is not the case for 

side-swipe, in so far as this type of accident is usually 

the result either of a loss of control, and an unsuccessful, 

too-delayed to avoid collision or else an error 

of judgment in distances when overtaking or at 

cro$sroads. 

Nurnber 

Figure 8. Truck lmpact direction 

692 


rrll, , ,.I, , ,r-az 

3 4 S s 7 i e t o 

Direction (clock face) 

r 6 a 

Direction 

To study the actual damage caused to passenger 

vehicles, we use a system of coding called VDI 

(Vehicle Damage lndex), the last parameter-crushing 

is used below, and is distributecl as follows: 

VDI Code no. 5 corresponds to damage reaching 

windscreen. Thus we see that our sample is mostly 

made up of seriously damaged vehicles. In fact, 

damage Code nos. 6-7-8,9 which correspond to some 

crushing of the survival $pace area, make up 50 

percent of our sample. 

As far as occupants are concerned, driver death rate 

is exceedingly high at 60 percent. We would point out, 

? t f 

(clock face)

. ! 

i 

tot 

: 

I 

Number flI 

tffis Lcl 

I 

roI 

I I 

oL- 

Flgure 9. Truck lmpact locatlon 

Number eo{ 

Figure 10. Pes$onger vehicle impact locataon 

6 

7 

B 

6 

Ntrmber a 

IFals Lct 

Figure 11. Trucldpassenger vehicle underrun 

2 

I 

o 


Right sftIc Stu 

Location 

Rllbt rldo Riil 

Location 

,_Nl , 

Ijft $i4c 

Slde-srtpc

18, 18X 

Number 


31, Bet 

Flgure 12. Truck/passenger vehlcle underrun (tl6) 

however, that only 20 percent of these drivers were 

restrained, a very low figure, even considering our 

study was carried out before the recent French Campaign 

to enforce seat-belt wearing. 

If we compare vehicle damage with occupant injury, 

we see that out of I I drivers with vehicle 

damage Code 6-7-8 or 9, only 2 survived (i-e 82 

percent death rate). 

These 2 cases merit further study. In the first one, 

the driver was restrained, and the VDI coded 8. This 

is a high rate of damage, but in this case, was limited 

widter since underrun was only one-thirds. The OAIST 

code for the driver is 3. 

In the second case, the driver was unrestrained, and 

the VDI also 8. However, this was a case of sideswipe, 

and since there was delayed impact. Furthermore, 

the passenger vehicle was not overrun by the 

truck: so the energy of the 2 vehicles was not totally 

used by the collision, which partly explains the OAIS 

of 2 (slight injuries). It would thus seem that the 

highest degree of damage is a criterion of impact 

severity, but which must be examined in association 

with underrun. 

If we observe the less-damaged vehicles, we find 

five deaths (50 percent). There is thus a correlation 

between VDI and injurity severity. Furthermore, for 

injured surviving drivers (6 drivers), there are 6 cases 

of side-swipe, thus no underrun. For passengers, there 

are no severe injuries, but the collisions in question all 

occurred with delayed impact on left-hand side, i-e, 

the cabin was not impacted on their side. 

lThe {)AIS ir the international coding system for occuPmt injury on a scale 

of 0 (uninjured) to 6 (deceased). 

694 

n Alt of fnont 

$l Tno-thinds 

fi 0ne-third 

I Side-swipe 

Fronto-Lateral Impact 

This type of impact involves 12 trucks and 12 

passengers cars (with 2l occupants). 

In this type of impact, underrun is less severe, and 

the impacting zone of the truck becomes a determining 

factor. Some zones, such as wheels, and spare 

wheels, $eem particularly aggressive, others less so, 

such as some types of underrun side-guards. Tractor 

fuel tanks, on the other hand, appear in our sample 

to be a relatively t'avourable zone: we have three cases 

in which the fuel tank undoubtedly absorbed the 

shock. It is also of interest that damage to passenger 

vehicles in this type of collision is clearly less severe 

than in the case of Fronto-Frontal impact. This is 

demonstrated in the following table, which indicates 

VDI distribution: maximum level is 4. 

t t t t t t t t t l 

I r l z | 3 | 4 | 5 I 6 | 7 | 8 1 9 | 

t t r t t t t t r r l 

l l 

I N"ofcaeee I I 13 | 5 | 3 | 0 | 0 | 0 I o lo I 

l l r t t l l l l l l 

In confirmation of this result, we find a lower 

occupant death rate (22 percent), even though only 

one driver was restrained. Similarly, amongst surviving 

occupants, we find only one injured person OAIS 

4, and 3 OAIS 3, which means that severely injured 

occupants also represent only 22 percent of total. As 

it is also the case with Fronto-Frontal impact, avoidance 

of underrun, when applicable, is a beneficial 

factor. 

Other Impact Types 

We have chosen to group together other types of 

shocks since our sample was only a small one. ln spite 

of the difficulty in the drawing conclusions for this 

category, we are able to point out that lateral impact 

is always severe for the passenger vehicle (as is also

the case for collisions betweEn 2 paswngers vehicles). 

Indeed, in our sample, death rate for occupants on 

impact side, when the cabin is damaged, is 75 percent. 

Survival rate for occupants on the opposite side 

depends on two factors: if they are restrained, and the 

speed of truck impact. 

As far as wearing a belt is concerned, our observations 

are drawn from studies carried out on collisions 

between passenBer vehicles, since none of the occupants 

in our survey were restrained. 

For the second point, as soon as truck speed can no 

longer be qualified as "slow" (or even "very slow"), 

we have no survivals. By slow speed, we mean speeds 

resulting from heavy braking (ground marks) or from 

an initial shock (which slows the truck considerably). 

Fronto-Rear impacts may be assimilated to Fronto- 

Frontal impacts with one notable difference: the 

speeds of the two vehicles are to be substracted rather 

than cumulated. 

Truck Occupnnts 

Only two truck occupants were injured. One, at the 

wheel of a tractor semi-trailer was in Fronto-Lateral 

collision with a CX, suffered a slight hand wound 

(OAIS l); the other sole occupant of a l0 t coach was 

hit at high speed by a R 25 in a Fronto-Frontal 

impact, and had a fractured lee (OAIS 2). This 

second case may be explained by the fact that in a 

coach, runlike other trucks, the driver's seat is relatively 

low. Furthermore, the bodywork of coaches is 

different to that of trucks, and more easily deformable. 

Conclusion 

Our study included 53 pairs of vehicles involved in 

accidents: this figure is obviously a very small one. 


But we would point out that, in 1,985, trucks involved 

in accidents resulting in corporal injuries represented 

only 5.5 percent of all vehicles involved. 

What may we deduce? First of all, it seems that the 

most frequent type of impact is the Fronto-Frontal; 

therefore this type will be the most interesting source 

of study for Truck/Passenger vehicle collisions. 

We have also seen that underrun is an important 

parameter in impact severity. our survey shows that 

the most frequent type of underrun is the two-thirds 

one. Indeed, this type represents almost one-third of 

all cases, one other third being complete underrun; 

the other configurations together making up the 

remaining third. 

Bibliography 

D. Vullin, Accidents avec sortie de chaussde en TPC 

sur autoroutes de liaison. 

B. Chretien, C. Danner, A. Ciesi, La constitution et 

I'evolution du Poids Lourd. Etude Seres-Onser 

I 980. 

LD. Neilson, R.N. Kenp, H.A. Wilkins, Accidents 

involving heavy goods vehicles in Creat Britainl 

frequencies and design aspects. TRRC Report 

474. 

M. Dejeammes, J.L. Masson, G. Vallet, M. Ramet, 

SdcuritC des vdhicules de transport utilitaires et 

commerciaux. ONSER. Aofit 1984. 

C. Uny, Accidents de circulation impliquant des Poids 

Lourds. Thdse prEsentde ir I'Universit€ Claude 

Bernard LYON. Novembre 1982. 

G. Cashera, Rdle des Poids Lourds en traumatologie 

routiere. au cours de collisions avec des voitures 

. ldgdres. Thbse prdsentde I I'Universitd Claude 

Bernard Lvon. 1981. 

Seat Belt Effectiveness for Heavy Truck Oecupants During a Collision 

Mitsuo Horii, 

Kunio Yamazaki, 

Japan Automobile Research lnstitute, Inc., 

Yuji Amemiya, 

Japan Automobile Manufacturers 

Association, Inc., 

Japan 

Abstract 

In spite of the abundance of passenger car seat belt 

studies in various related fields, seat belt tests and 

studies for heavy trucks, which are substantially 

different in body structure from passenger cars, have 

been the minority. 

To obtain basic data for evaluating the seat belt 

effectiveness of heavy trucksn the following were 

studied: 

l) The analysis of prediction of ratios by which 

the effectiveness of seat belt$, based on a 

statistical study of accidents involving heavy 

trucks. 

2) Sled tests were conducted by simulating the 

barrier impacts of heavy trucks, and the 

magnitude of impact on the occupant wearing 

seat belt$ and not wearing them was 

compared. 

695

3l And the tests rryere conducted with different 

sled floor acceleration, impact speed, and 

their influence on occupant behavior was 

studied. 

The results of the above l) - 3), and their analysis 

results are reported as the first step in safety study for 

heavy truck occupants. 


Many studies have been made so far on the seat 

belts, those of passenger cars in particular, by leading 

nations of the world. These studies cover a wide 

scope, including (1) the evaluation of seat belt effects 

based on accident survey results, (2) the development 

of seat belts designed to improve occupant protection 

characteristics, and (3) the development of passive 

seat belts. 

On the other hand, many nations have made 

statistical studies of accidents involving heavy trucks, 

and after analyzing the study findings, have reported 

that the following are important for the heavy truck 

safety(l)-(7) 

l) The necessity of seat belts 

2) Reinforcement of cab strength to seeirre 

survival space 

3) Rollover countermeasures 

4) Prevention o[ under-ride in front and rear 

5) Improvement ol energy absorption by the 

steering wheel and dashboard. 

In the case of experiments concerning safety during 

heavy truck collisions, reports and data are primarily 

related to 2) and 4) above; that is, the reinforcement 

of cab strength and the prevention of under-ride(g)- 

(10). 

As for studies of the seat belts of heavy trucks 

where body construction, dimensions, specifications, 

etc. differ from those of a passenger car, we can cite 

a single example made in 1980 in West Ciermany at 

the Battle Institute,(ll) in which they studied the 

performance of passenger protection under various 

belt conditions. 

In spite of clamorous appeals for the necessity of 

seat belts, studies dealing with the seat belts of heavy 

trucks are seldom found among survey analysis and 

the results of heavy truck accidents, 

To investigate the effectiveness of seat belts when a 

cab-over type heavy truck (GVW, 20 tons) has a 

collision, we made a systematic study of the scale of 

impact on the occupant and the scale of injury to the 

occupant in relation to iftpact speeds using and not 

using seat belts through (l) the analysis of prediction 

of ratios by which the effectiveness of seat belts. 

based on a statistical study of accidents involving 

heavy trucks are shown, (2) sled tests under various 

696 


seat belt systems and (3) a full scale collision re$r as a 

preliminary test. Reports thereof follow below. 

In the sled test, experiments were made so as to 

generate a nearly equal impact to that of a heavy 

truck when it makes a frontal collision with a flat 

barrier. To establish the impact conclitions for the sled 

te$t, a test of fronlal collisions of heavy trucks with 

the flat barrier was made first to determine the scale 

of impact on rhe body during collision. 

On the Effectiveness 

of Heavy Truck 

Seat Belts Seen From Accident Survey 

Findings 

Figure I shows the heavy truck accidents (9,710 

cases) involving heavy trucks in 19g3 in Japan by 

collision areas. Accidents involving pedestrians and 

motorcycles are excluded. 

The front is the overwhelmingly dominant collision 

area with 72-ZTo of fhe total suggesting a considerable 

level of seat belt effectiveness as long as a survival 

space has been secured in the cabin. 

Figure 2 shows the occurrence ratio of the part of 

total accidents which consists of accidents where the 

cab front was hit and single accidents (roll-overs, 

collision with structures, etc.) where occupant injury 

levels would have been lowered by seat belts. 

The hatched area in the graph shows the number of 

drivers killed or seriously injured in spite of little cab 

deformation (that is, the $urvival space in each cabin 

was secured) in the rear-end collisions. frontal colli_ 

sions and single accidents. In other words. this data 

shows that seat belts would have alleviated the injuries 

suffered by this number of drivers. 

In the case of heavy truck accidents involving 

deaths and serious injuries, it is predicted that such 

injuries would be alleviated by about ZSt/o if secondary 

impact in the cabin is eased, or prevented, by 

using seat belts. 

The number of the dead or seriously injured in the 

graph (78 persons) are for the drivers alone and the 

number will increase if the passengers are inclucled. 

-e 

o 

--} 

IN 

)[l 

I l 

I t 

Lt I 

? 

(i) FrunLrr (:,,r I tsron 

O stde colttrrm* 

O R€ar-end col1trtfi 

Note) i The side cdllr*ron, t7.lX 

HB c,l4xlar€d froE rhe tsL!htlon 

of sccldeilre ciluicd bi 

sudden budplng ilnd dlrinfl 

IeIL/, irhr trrns, 16gy iadlud., 

.iol.llslons wlrh hegvy 

truck froni+, 

Figure l. Heavy truck accident$ by collision areas (ln 

Japan, during 19BB)

+ 

The numler of drivert serlously injured 

or killed in epite of little cab 

defornati on, 

Thc number of drivers seriouslv injured or killcd slen 

cab were seriously damaged. 

Figure 2. Prediction of seat belt effectlyenesa In 

heavy truck accldents 

As seen, a rough picture of the degree of the 

effectiveness of seat belts in heavy truck accidents was 

envisaged from ascident survey findings. Though 

accident survey$ can clarify statistical ratios, they 

cannot explain how impact velocity and degrees of 

acceleration affect drivers' injuries. Hence, full scale 

collision tests and sled tests were made to draw 

conclusions concerning these effects. 

A report of these tests follows: 

Frontal Collission Test of a Heaw 

Truck With a Flat Barrier 

To establish impact conditions for a sled test 

required to review the effectiveness of seat belts when 

a heavy truck makes a frontal collision with the 

barrier and to investigate the passenger behavior 

during impact condition, a preliminary frontal barrier 

impact test of a heavy truck was made to study the 

scale of impact applied to the cabin, the duration 

time. etc. 

A special emphasis was put on the degree of cabin 

floor acceleration which greatly affects an occupant's 

injury level in a collision. 

ln the current test, the collision speed was determined 

to be 20 mph (32 km/h) based on foreign 

studies(5),(ll),(12) made so far on 

(l) Driving speeds 

(2) The composition of impact speeds in heavy 

truck accidents and 

(3) Impact speeds used in actual test. 

In addition, to study dummy behavior in a collision, 

three Pan 572 dummies (HYBRID-II) were put 

on the driver's seat, the center seat, and the passenger's 

seat with a 3-point seat belt (NLR), a 2-point 

seat belt, and a no belt, respectively. A measurement 

of injury values was also carried out. 

Figure 3 shows a test scene. To examine dummy 

behavior during a collision, the door panel was cut 

open for the test but it was reinforced to provide the 

same Iongitudinal crush strength of the door in a 

collision as before the cutting (the standard condition). 


'$s{rM.r,ld: .m;ffi--Yr 

Figure 3. A test scene 

[Test Results] 

Figure 4 shows the deformation around the cabin, 

after the test. 

i) Cabin deformation 

As the main frame is highly rigid, the deformation 

of the cabin after collision is slight, roughly 6090 of 

the passenger car level if absolute values of deformation 

are cornpared. The crush stroke of cab frontpanel 

is approximately 180 mtn and the survival space 

in the cabin is secured sufficiently. 

ii) Acceteration of the Cabin 

The waveforms of cabin floor acceleration, measured 

in the neighborhood of the driver's seat and the 

cFt' 

# ffi 

Figure 4. Cab deformation after the test

Gso 

E 

P 

L 

(u 

G) 

(d 

L 

o 

q- 

€ 

40 

30 

20 

T'ime (ms ) 

Figure 5. Waveform of cabin floor acceleration 

passenger's seat, have two peaks of duration time of 

acceleration of 40 to 45 ms and the maximum 

acceleration of 55 to 60C as shown in Figure 5. As a 

whole, the waveform can be patternized as a triangle 

waYe, 

Compared with those of passenger cars, this waveform 

features a short duration time, roughly 50Vo 

shorter, and a steep rise. 

Irrespective of the occupant's use of a 2-point seat 

belt or no belt, ride-down effects are not usually 

realized unless the duration time is longer than 100 

ms. Therefore, in the case of an impact where l.he 

duration time is 40 to 45 ms at best, as in this test, 

the occuparts are caused secondary impact with 

interior $tructures and devices at the initial speed 

maintained before the collision aftcr the vehicle is 

stopped without re$pect to the shapes of the acceleration 

waveforms of the cabin. This short time waveform 

therefore implies critical conditions for the 

driver. 

iii) Dummy behavior 

Figure 6 shows the behavior of a dummy during 

impact condition by the time series. 

r Driver's seat (3-point belt) 

The knees come in contact with the lower 

part of the instrument panel as the belt 

extcnds. However, other parts do not come 

in contact with the panel. 

r Center seat (2-point belt) 

The torso fell forward with its hip in position 

to allow the head and face to hit the 

instrumint panel. 

r Passenger's seat (without belt) 

After the knees hit the instrument panel, the 

EXFERIMENTAL SAFETY VEHICI,ES 

torso fell forward while the hip was raised 

upward, and the head went out through the 

front glass to reach the barrier. 

Dri ve r' s 

sea t 

( 3- po'i nt be I 

( 2-point be 

Displacement (mm) 

ror 

'ii", 

xlo 

1' 

ao 

rol,;o9-P- 

Displacement (mm) 

Since valrres of shoulder, hip and knees 

ctruld not read as they were behind the 

dunrnry seatinq in the pdssenger's seat, 

values for the head alone are shovrn. 

Passen- 

Di spl acemettt (ntm) 

tn- -*t -" 

,oo ll 

f i 

w-/ 

/ 

!rro'./' 

t++:j 

t i 

6 0 

t?o 

r 

.,nJ ; 

f - F t - . - - 

Flgure 6. Dummy behavior during impact conditlon 

/' 

{* 

o 

6 

a 

I 

a.) 

o 

0 

6 

o 

E 

o 

E 

U 

6 

6 

o

iv) Values of passenger injury 

Figure 7 shows thc injury values obtained from this 

test as reference value. 

r) 

2) 

3) 

If a 3-point belt is wornn impact on the head, 

the chest and the femur are all below the 

injury criteria of the human tolerance prescribed 

bv FMVSS 208, but 

If a 2-point belt is worn or no belt, the head 

comes in contact with the instrument panel 

or with the barrier front, the head penetrates 

into the windshield, respectivcly, pushing up 

head injury values much higher than 

HIC: 1000. 

As for the chest and the femur, injury values 

are less than the injury criteria of FMVSS 

208 irrespective of the use of a belt. 

3-point bel t 

2-point bsl t 

No belt 

3-l)oint bel t 

?-poi nt be1 t 

No belt 

3-point belt 

2-point bel t 

No bel t 

Figure 7. Passenger Inlury valuee (trontal collision of 

a heavy truck wlth the flat barrier at 32 

km/h) 

Sled Test 

Injury criteria of Fl.iVSS 208 

+, 

Che, i- dcceleration (G) 

Femur load (kgf) 

Outline 

To study the effectiveness of a seat belt when a 

heavy truck makes a collision, the HYGE impact test 

equipment, hereafter called "the sled", was generates 

when it makes a frontal collision with the flat barrier. 

In the test, a device modelled on the compartment of 

the driver's and passenger's seats of a cab-over tyFre 

heavy truck was fixed on the sled and a given level of 

impact was administered to it after loading belted 

dummies or a beltless dummy on it 

Figures 8 to l0 show the test conditions. 

Part 572 dummies were used as human body 

dummies and seat belts were fastened in the following 

manner: 


(1) Driver's seat 1) 3-point belt with NLR 

2) z-point belt with NLR 

3) No belt 

(2) Assistant's seat 1) 3-point belt with ELR 

(that $enses floor acceleration 

of 0.4G) 

2) z-point belt wirh NLR 

3) No belt 

Figure 8. A eled test scene 

Figure 9. Driver's seat (3-point belt with NLR in us6l 

Figure 10. Assistant'sseat 

(3-point belt with ELR in 

use)

Impact Conditions 

Generally, the duration time of acceleration applied 

on the body floor in a frontal collision with the 

barrier remains almost constant irrespective of the 

impact speed and, in the case of a heavy truck, is 

roughly 50 ms. 

In the sled test, the duration time of the floor 

acceleration of the sled was set constantly at around 

50 ms to give a similar impact to that generated by 

the heavy truck when it makes a frontal collision with 

the flat barrier and impact were given under various 

acceleration degrees, that is, different impact speeds. 

Test Results and Review of Results 

l) Comparison of the results of a full-scale 

collision test and of a sled test 

Figure 12 compares the waveform of cabin floor 

acceleration in the full scale frontal collision test (32 

km/h) with that of the sled floor acceleration in the 

sled test (at 37 km/h). In the case of the full scale 

test, the waveform shows a very steep rise but, in the 

case of the sled test, it is as if the waveform of the 

former is simply patternized into a triangle wave, or a 

half sine wave. 

Figure 13 compares the full scale test results with 

the sled test results in terms of injury values incurred 

at different sections of the dummy seated on the 

driver's seat with a 3-point belt. 

The black round marks in the drawings indicate the 

results of the full scale test. The sled test results 

relatively correspond to the full scale test results. This 

means that the behaviors of dummies correspond in 

both tests. 

2l Dummy behavior 

Figure 14 shows the dummy behavior during impact 

in the passenger's seat by the time series. 

When a 3-point belt was worn, neither the head nor 

the chest come in contact with the instrument panel. 

In the current test, however, the knees come in 

contact with the lower part of the instrument panel if 

the speed exceeded 35 km/h, restricting the knees 

behavior. 

1 

Time: Roughly 50 ms, constant 

Acc.: Varied between l0 and 45G 

Flgure 11. The waveforms of floor acceleration of the 

sled 


(5 

E 

o 

+) 

lg 

L 

(u 

c) 

(J 

(J 

rC 

L 

o 

|! 

Figure 12. Gomparison of the acceleration waveform$ 

In the full scale teet and the sled teet 

If a 2-point belt was used, the torso fell forward 

with its pelvis in position to allow the head to hir the 

instrument panel top. 

In the test, the speed ranged between l7 and 42 

km/h, but a certain area of the instrument panel top 

was hit by the head without exception. If the capacity 

of impact energy absorption of this area is improved, 

it is possible to reduce the impact on the head. 

If no belt was used, the dummy moved fbrward in 

its initial position at flirst, hit the lower part of the 

instrument panel with the knees, then hit the instrument 

panel front with the chest. Then, while the 

pelvis was raised upward, the head hit the instrument 

panel top, and then hit the windshield. Thc dummy 

was not thrown out of the vehicle in the test a$ a 

preventive net wa$ fixed on the windshield. In an 

actual accident, however, it is likely that the passenger 

would be thrown out. 

3) Head injury Criteria (HIC) 

FrontaJ co'l'l'ision 

of a 

l0-ton truck with the 

bamier at 32 km/h 

(20 mph) 

Sled test at 

37 kn/h (23 mph) 

L, 50 

Time (ms) 

As an example, Figure 15 shows HIC changes in 

impact speeds derived from the test results with and 

without a belt. 

The straight and dotted lines in Figure 15 show the 

driver's and the passenger's value, respectively. 

HIC:1000 is reached around 35 to 45 km/h when 

either a 2-point belt is u$ed or no belt, while it is 

reached at around 55 km/h* (on the passenger's 

value) when a 3-point belt is used. 

If these are evaluated by the impact speed at which 

a heavy truck makes a frontal collision with the flat 

barrier, HIC: 1000 will hypothetically be reached 

around 30 to 40 km/h if a 2-point belt or no belt is 

used and around 50 km/h* (on the passenger'$ $eat 

position) if a 3-point belt is worn.

d 

s 

a 

C 

E 

€ o 

E o 

o 

r) Speed vs, IIIC 

?00 

{00 

Col I i sion or i|npdct speed 

0 ro (km/h) 

G00 

lo(J 

t000 

Full scale tes{ 

2) Speed v6. che8E Acc. 

0 

?0 

t0 

i0 

l0 ?0 30 10 

t 

. , / 

Ful I scale tes! 

jled Iest 

Sled tert' 

3) Speed vg, shoulder belE load 

0 

r0 ?0 30 t0 

L O 

E'J 

B€ 

,EO 

n 

a(t 

?n0 

t00 

$00 

t00 

t0 00 

lr 00 

4) Speed vs, pelvls acc. 

0 

r00 

?00 

Full scale test 

Full scale test 


SI ed test 

Sl ed test 

.t' 

SS klll/h 

Figure 13. Comparison ol test results between full 

scale test and sled tests 

Note: The corresponding speed on the driver's seat 

position to the asterisked speed is predicted 

to be over 60 km,zh but, as the degree of 

extrapolation is large enough as shown in 

Figure l5 l), the evaluation was made on the 

passenger's seat position. 

The report thus far has been a prediction of 

marginal speeds based on test results and only shows 

the results of an example. lf whether or not this 

represents the general tcndencies of a heavy truck 

should be judged after examining the results of many 

tests to be rniicle in the futurc. The samc applies to the 

chest injury value, etc. explained to the right. 

[.]-point BeltJ 

[?-point Belt] 

[No Bel t] 

Displ.rctlrntnt (rtm) 

iE_r-il 4..4 ry 4,f,j_Q__|4_!S0 

0isFl acement (rrya) 

Figure 14, Dummy behavlor durlng lmpacl 

4) Chest acceleration 

- 26 kn/h 

-"- 37 km/h 

-.- 4? krn,/h 

- l7 k$/h 

.--- 28 kn/h 

*.- 39 krn/h 

Figure 16 shows the relation between impact speeds 

and chest injury values (3 ms G). 

For the same impact speed, the injury value on the 

driver's seat is different from that on the passenger's 

seat but the impact speed at which the injury criteria 

of FMVSS-208 of the chest of 60G (3 rns cut G) is 

reached is roughly 35 kmlh if no belt is used and 

roughly 45 to 65 km/h if a 2-point or 3-point belt is 

used. This clearly suggests the use of a belt is 

effective. 

5) Femur load 

Figure 17 shows changes in the femur load by 

impact speed when a belt is used and when it istt't. 

While the femur load remains less than j of the 

limit value within the scope of the current test only if 

a belt is worn l) and 2) of Figure 17, the human 

701

Figure 15. HIC changes in lmpact speeds 

EXPERIMENTAL SAFETY VEHICI.ES 

Flgure 16. Changes in the chest acceleration by impact speeds 

702 

- 

o 

o 

4 

O 

60 

4h 

I ) 3-point bel t 2-point bel t 3) f{o bel t 

20 40 

Irrpact speed V (km/h) 

r< Driver with 3,point 

bel t NLR 

tJs* = Z3o 

g.---g Assistant with 3-point 

bEIt ELR 

0s = 60" 

3000 

?000 

r000 

20 40 60 

Impact speed V (kn/h) 

.€ Oriver with Z-point belt NLR 

(Head comes jn contact with 

steering r{heel ) 

EI-..€ Assistant with ?-pojnt belt 

NLR 

(Head comes 

' 

in contact with 

instrument panel ) 

| / J-polnt Dett ?) z-point bel t 3) No belt 

r- 

?0 40 60 

Irnpact speed V (km/h) 

Driver with 3-point belt 

lj!^= zs", e llpHJfi[noEJl, - 

tiAn 

'i0 

mm Up 

El---'E Assistant witn i-pornt 

bett ELR 

!s = 60''' 

20 40 60 

Impact speed V (kn/h) 

........-[)river with Z-point belt NLR 

Chest comes in contact with 

steering wheel 

El---€ Assistant with ?-point belt NLR 

The head cJmos ifl contact with 

'instrument 

panel but the chest 

. does not as il lustrnted. 

= 

3000 

2000 

rm0 

?0 rtO 

Impdct speed U (km/h) 

r+ Driver with non belt 

(Head comes in contact 

with steering whee'l ) 

E!-.-tr Assistdnt with non 

bel t 

(Head comes in contact 

wj th j nstrument panel ) 

20 40 

Impact speed V (km/h) 

r:_ 

Driver with flo belt 

Chest comes in contact 

wrth steering wheel 

d-r-fl Arsistafrl rdith no belt 

Chest comes in contact 

with instrument panel


l) 3-point belt Z) 2-point belt 3) No belt 

_ _l iai.t .v.a. j gq - -_l=irnl.t_ _va].ue 

20 40 60 

Impact speed V (kmi 

h) 

o 

E 6,m 

L 

E 

'P 

....._ Priyel.wjth 3-pojnt 

DEII NLR 

B---g Assistant wjth 3-pornr 

bEIC LLR 

Figure 17. Ghanges in femur load by impact speeds 

tolerable level (1,021 kg) is already reached at an 

impact speed of roughly 35 km/h, if no belt is used 3) 

of Figure 17, as the dummy's knees come directly in 

contact with the instrument panel, clearly suggesting 

the effectiveness of seat belts, 

6) Occupant injury values by impact speeds 

So far, occupant injury values by impact speeds in 

the sled test have been analyzed. In Figure 18, the 

above speeds are sonverted into the impact speeds at 

which a heavy truck makes frontal collisions with the 

barrier based on the results of 

Plotted speeds in the graph show the limit speeds at 

which the injury criteria prescribed by FMVSS-208 are 

deemed to have been reached. 

As the conditions of the driver's seat are different 

from those of the pas$enger'$ seat with respect to the 

following, comparison cannot be made directly. 

l) Seat belt anchorage position. 

2) Retractor type (NLR for driver, ELR for 

Passenger). 

3) Secondary impact position of passenger and 

respective absorption of energy. 

4) Alloted rate of loads on the shoulder and lap 

belts, etc. 

If seat belt effectiveness is reviewed in general, 

one 3-point belts, two 2-point belts and three no 

belts are effective in protecting the passenger, in that 

order. 

The limit speeds on ihe driver's seat position when 

a belt is used and when it isn't are as follows: 

?0 40 60 

lmpact speed V (kmlh) 

.+ Dr-iver with z-point 

bCIt NLR 

g---E Assistant with ?-point 

bel t NLR 

o 

d f,UIJ 

o 

L 

o 

20 s0 

lmpact speed V (kffi/h 

+ . DriYea si15 

n0 0elt 

E--€ Ass i stant wj th 

no bclt 

Non belt , at around 25 to 30 km/h, the 

chest or femur will reach its 

limit 

z'point belt at around 40 km/h, the head 

will reach its limit 

$-potnr oerr . at around 5O km/h, the che$t 

will most likely reach injury 

criteria prescribed by 

FMVSS-208. 

uriver't 

tent 

ir5 

5ea 

section Llnlt speed (km/h) 

ilo bel t Hend - 

2 -potnt bel t 

Chest 

Head 

_*rjlg:l_ 

l-F jn t bel t hen0 (E 

ttcrt - 

cffi Thc dbove.e" 

sults 5ulqest the 

ljkclinesr nf a higher 

I inj t rt,red thah tne 

Figure 18. The collision speeds at whlch the human 

tolerance (iniury eriteria prescribed hy 

FMVSS) are reached 

_. 

-r- 

,{o belt Hea 0 t-I 

lhe5 t 

- 

?-Folnt bel t 

chest 

3-point belt Head - 

(;hd:it - 

Fcri ur 

IIE 

- 

- 

{E:f 

dd 

703

Similarly, limit 

tion are: 

speeds on the assistant's seat posi- 

Non belt . at around 25 to 30 km/h, the 

chest will reach its limit 

z-point belt. at around 30 to 35 km/h, the 

head will reach its limit. 

3-polnt belt. at around 35 to 40 km/h, the 

chest will reach its limit. 

7') Occupant injury values by seat belt fastening 

angles 

Figure 19 shows the relationship between HIC and 

the impact speed of a 3-point ELR on the passenger's 

seat by using shoulder belt fastening angles as param_ 

eters. 

ln the test, the angle of belt fastening (ds) was 

changed from 26" to 60" but HIC values changed 

little if impact speed remained the same. 

Figure 20 shows a similar relationship with respect 

to the acceleration of the chest. ln this ca.se, the chest 

injury values tend to be irrfluenced by shoulder belt 

angles. 

Conclusion 

The findings of a survey of the effectiveness of a 

heavy truck's seat belts derived from accident survev 

results, the results of full-scale heavy truck collision 

tests, and the results of sled tests made to review the 

effectiveness of a heavy truck's seat belts explained so 

far can be summarized as follows: 

' 

l) The effectiveness of a heavy ftuck's seat belt 

derived from accident survey results 

As a result of investigations into the effectiveness of 

a heavy truck's seat belts by using fhe statistical data 

of heavy truck accidents that occurred in Japan 

during 1983, it is reasonable to predict that the 

number of fatalities and seriously injuries could be 

-4/ 

F 

I 

J 

fl 

o 

I 

o J 

I J 

d 

9s 

r: ?6o 

o,38" 

D , 60o 

10 l0 30 40 50 60 

Impact speed V (km/h) 

tj 

Figure 19. lmpact speed VS HlC, where the angle of 

the shoulder belt fastening is used as a 

parameter 


-fo 

6o 

E a.o 

qJ 

Tro 

U 

Ez0 

(u 

Figure 20. lmpact speed vs. chest acceleration; where 

the angle of the shoulder belt fastenlng is 

used as a parameter 

reduced by roughly Zjolo if occupants 

of heaw trucks 

were wearing seat belts. 

2l A full-scale heavy truck collision test 

As a result of a front collision test with the flat 

barrier of a cab-over type truck in the GVW 20_ton 

class at an impact speed of 32 km/h, it was discov_ 

ered that 

l. The deformation of each section of cabin was 

minimal and the cabin's survival space after tests was 

secured sufficiently. 

2. The acceleration of the cabin floor, which is 

reproduced as the floor acceleration of the sled test, 

featured a waveform where the duration time was 

roughly 5090 shorter and the rise was steeper than 

that of a passenger car. This waveform is critical in 

that it suggests secondary collisions of passengers. 

3. While the injury values of the head, chest, and 

femur do not reach the human tolerance prescribed 

by 

FMVSS 208 if a 3-point belr is in use, the correspond_ 

ine HICs of the use of Z-point belts and no use of 

belts are well over 1000. 

3) Sled tesrs 

! 

c,, 

"r,,1,{ 

i'/ 

t0 ^-^ 0S = 260 

0s = 38" 

o*-+ 0s = 60" 

10 20 30 40 50 60 

Inrpact speed ! 

( kn/h ) 

I. By the sled tests, a dummy behavior similar to 

that in the collision of a heavy truck with the barrier 

was reproduced. 

2. As a result of sled tests similar, it is clear that 

the use of belts is effective for passenger protection. 

On the driver's seat, an impact speed corresponding 

to the human tolerance prescribecl 

by FMVSS 20g is 

predicted as follows: 

25 km/h or so if no belt is used. 

40 km/h or so if a 2-point belt is used.

50 km/h or so if a 3-point belt is used. 

3. In the test, attempts were made to identify the 

influence of the angle of the shoulder belt fastening 

(ds) on the pa$senger injury value. As a result, it was 

found that the HIC had almost nothing to do with the 

value of ds, but chest injury value tend to be 

influenced by shoulder belt angles. 

Recommendation 

This time. we made various tests and studies to 

examine the effectiveness of the $eat belt as part of 

our study of a heavy truck's occupant protection and 

could clarify its effectiveness to some extent. The 

following, however, are conceivable future problems, 

l) When a 2-point belt is used, the head and chest 

comes in contact with the steering wheel in the 

driver's seat and the head hits against the instrument 

panel in the passenger's seat, making damage more 

serious. The energy absorption by the steering wheel 

and instrument panel during secondary collision will 

have to be clarified further from the viewpoint of 

occupant protection. 

2) When an ELR-type seat belt is installed in a 

heavy truck, attention should be directecl to the 

following: 

The range of standard acceleration sensed by the 

current retractor locking mechanism should be set 

between 0.45 and 0.8-5 by each country in the case of 

an acceleration sensing formula for the vehicle body, 

But the level of sensed acceleratiorr varies from 

country to country, and Japan which approved 1.5 C 

products for automobiles produced after September 

I 987. 

In addition to a study of the inconveniences caused 

by seat belts, including excessive constriction of the 

stomach during rough road driving, a reasonable 

standard acceleration value should be agreed to internationally 

in the future. 

3) In this test, rigid seats without suspension were 

used, but the use of suspended seats is possible for a 

heavy truck and suspension seats will also have to be 

examined in the future. 

Acknowledgement 

Lastly, we would like to express our deep appreciation 

for the cooperation of JAMA's rnet'ubers in the 


heavy truck study committee and the heavy truck cab 

impact W/C committee 

during this study. 

Reference 

l. Henry E. Seiff, Heavy Truck Safety-What we 

know, Tenth International Technical Conference 

on Experiment Safety Vehicles, 1986. 

2. Michigan Univ., Truck Involved in Fatal Accident, 

l98l PB84-183037, 1984. 

3. Robert M. Clarke, Joseph Mergel, Heavy Truck 

Occupant Crash Protection. A Plan for Investigatirrg 

Ways to lmprove It. SAE paper 821270. 

4. K. Hogstr6m, L. Senenson(s), Injuries in Heavy 

Trucks and the Effectiveness of Seat Belts. 

VDl-Berichte Nr. 368, 1980. 

5. L. Langwieder, M. Danner, M. Wrobel, Ein 

Beitrag Zur Risikoanalyse Und Characteristik 

Von LKW-Unfiillen. SAE paper 845017, 1984. 

6. H. Biirger (D), Bedeutung und Rangfolge von 

Sicherheitsbnahmen am Lastkraftwagen. VDI- 

Berichte Nr. 368. 1980. 

7. F.L. Krall and G.W. Rossow, Heavy Truck 

Safety...The Need to Know. Traffic Quarterly, 

vol. 35, No. 3, 1981. 

8. B.S. Riley, S. Penoyre, H.J. Bates, Protecting 

Car Occupants and Cyclists in Accidents Involving 

Heavy Goods Vehicles by using Front Underrun 

Bumpers and Sideguards. The Tenth International 

Technical Conference on Experimental 

Safety Vehicles. 1985. 

9. A. Yamanaka, M. Sato, N. Nagaike, and T. 

Nishida, The Research and Development of Nuclear 

Fuel Safety. Transporter and its Evaluatiorr. 

XXI Fisita Congress, June, 1986. 

10. H. Biirger, Dynamische und Statiche Untersuchung 

der Kollisions-Festigkeit von LastkraftwagenFaherehdusern. 

Automobiltech Z vol. 83, 

No. 3. 1981. 

ll. C. Rtitter, H. Honschik (D), Sicherheitsgurte in 

Nutzfahrzeugen, XVIII International Congress, 

1980. 

t2. E. Franchini, Truck Crush Testing, SAE paper 

700411, 1970. 

'705

Front Undernrn Guards for Trueks 

B.S. Riley, 

A.J. Farwell. 

T.M. Burgess 

Department of Transport, 

United Kingdom 

Abstract 

This paper presents research which is a continuation 

of earlier work on front underrun guards for trucks, 

carried out at the Transport and Road Research 

Laboratory in Great Britain. 

Results are given from impact tests between European 

small cars and trucks, where both vehicles are 

moving towards each other. Comparisons are made 

with earlier tests at similar closing speeds but where 

the truck was stationary belbre impact. It is concluded 

that there is little difference in the results from 

either method of test, particularly for the case when 

an underrun guard is fitted. 

The results suggest that an energy absorbing truck 

front underrun guard with a ground clearance of 300 

mm is able to provide protection from fatal or severe 

injuries to seat belted occupants in a small car (750 

kg) at closing speeds up to about 65 km/h. 

lmpact tests have been carried out between a rigid 

faced 250 kg trolley and three different types of 

energy absorbing truck underrun guard. The test 

procedure thus developed could form the basis of a 

legislative test to determine whether front underrun 

guards have adequate energy absorption. Limits are 

suggested for the proposed test procedure to cover 

speed of impact, input energy, size of trolley face and 

trolley decelerations during impact. 


This paper describes work carried out at the Transport 

and Road Research Laboratory in Great Britain 

and is a continuation of research presented at earlier 

Experimental Safety Vehicle Conferences in 1980t and 

19852. On this occasion it is concerned only with 

underrun guards fitted to the fronts of heavy goods 

vehicles to reduce the severity of injury to car 

occupants. 

Accident data 

Table I shows the number of car occupant$ killed 

each year in GB in accidents involving goods vehicles 

of gross weight over 7.5 tonnes for the period 1974 to 

1985. The number of car occupants killed annually in 

all other road accidents is also shown. 

There was a general decline in the numbers of all 

categories of road user killed in truck accidents until 

around 1980. This may possibly be attributed to the 

gradual diversion of heavy goods vehicle traffic to the 

706 

EXPERIMENTAL SAFETY VEHICI.ES 

safer motorway roads in this period. Since 1980 the 

number of occupants killed each year has levelled off 

to around 320. Approximately two-thirds of this 

number are killed when their vehicles impact the 

fronts of trucks, most frequently when the two 

vehicles are travelling in opposite directions and the 

front of the car hits the front of the truck. 

Heavy goods vehicles are very aggressive toward$ 

cars in collisions for several reasons. The large mass 

ratio results in the velocity change of the car being 

much greater than that of the truck. The height of the 

truck structure is such that, in a collision, the car may 

run under the truck, often to the extent that the truck 

structure comes into contact with the car occupant 

compartment or in extreme cases into direct contact 

with the occupant. By this mechanism, the important 

energy absorbing zones of a car, which tend to be 

below the truck structure, are not used to protect the 

occupant. Finally, because of the rigidity of the truck 

structure, most of the impact energy is dissipated in 

the car rather than in the truck. 

Truck front underrun guard design 

The first essential step in providing protection for 

car occupants is to lower the front structure of the 

truck so that effective use may be made of the energy 

absorbing structure of the car. It is also desirable to 

make the lowered structure of the truck energy 

absorbing. This has the effect of increasing the 

maximum survivable closing speed for the car occupants, 

which is particularly beneficial in car front to 

truck front impacts where closing speeds tend to be 

higher. 

The amount of energy absorption that may be 

provided by the truck frontal structure as it yields is 

limited by two factors. Firstly, the total longitudinal 

crush of both the car and truck structure must not 

exceed the original length of the car bonnet. Secondly, 

the force level at which the truck structure yields must 

be compatible with the tbrces at which a car front 

collapses so that at the end of a severe impact both 

have crushed. This latter point is important. If a 

guard or other form of protection is very soft, it will 

Table 1. Gar occupants kllled In road accldents In 

Great Britain. 

YEar 1974 1976 1978 1980 1 981 1982 1t8.t 1 984 1985 

Acc idents 

irrvol v irrg 

HGVs 

All 6lher 

ecc iderrts 

471 ]91 It7 296 12J 115 J00 J'i2 125 

2238 2'129 2272 1982 1964 2108 17't9 1845 1116 

Iotal 2109 2520 7569 2278 2287 2441 2n19 2197 2061

just bottom out and absorb little energy. Similarly, if 

it is too stiff (for a snrall weak car) it will not yield 

and again will not absorb rnuch energy. 

One effective method of increasing the amount of 

possible energy absorption would be to bring the face 

of the guard forward of the truck front. An increase 

in stroke, even as little as 200 mm, introduced in this 

way would make a sub$tantial difference to the energy 

absorbed. Changes in the measurement of vehicle 

length for legislative purposes would be needed to 

allow the many vehicles already operating at maximum 

length to make use of this concept. 

The limiting stroke and force levels that can be used 

at present suggest that impacts between small cars and 

trucks provide the important design case for the 

required truck structure. 

fo, ffi Wfl r,t ' tt 

$;+ 


The results from this earlier programme of work 

suggested that energy absorbing truck guards should 

ol'fer protection to seat belted car occupants at closing 

speeds up to 65 km/h. lt was also considered that 

grourtd clearance should not be much greater than 300 

mm and certainly no more than 400 mm, otherwise 

the structural parts of smaller cars would be overridden. 

It was estimated in Reference 2 that with this order 

of protection it should be possible to prevent at least 

60 car occupant deaths each year in Great Britain. 

Objectives of current work 

All the earlier impact tests have been carried out 

with the truck stationary. In a study of fatal road 

accidents involving trucksa it was observed that when 

the truck was moving forward at the time of initial 

impact, it tended to ride up onto the car as the latter 

was pushed backwards. At the end of the main impact 

Previous TRRL work on front underrun 

guards' 

The TRRL research on this subject which was the truck thcn settled down onto the car cflusing 

presented at the 1985 Experimental Safety Vehicle further distortion of the passenger compartlnent. 

Conferencez, was based on a series of test impacts Therefore one objective of the present research has 

between the fronts of cars of various sizes and the been to investigate the effect on the protection offered 

front of a truck fitted with an energy absorbing front to the car occupant when the truck is moving. To this 

underrun guard. In these test$ the truck was station- end, impacts have been carried out between trucks 

ary and the car was travelling illto it at about 65 

km/h. The guard used was one developed jointly 

and small cars with both vehicles moving towards 

each other at the closing speeds that were used in the 

between TRRL and TI Tube Products Ltd and its earlier programrne of work. The tests have been 

energ,y absorption is based on the plastic deformation carried out with trucks both fitted and not fitted with 

of mild steel in the form of two invertubes. The guard a front guard. 

may be seen in Figure l. Each invertube collapses at a The second objective is coucerned with the capabil- 

near constant load of just under 50 kN over a stroke ity of a truck front underrun guard to absorb energy. 

of 200 mm giving a total energy absorption of about If legislation is introduced to require trucks to be 

20 kJ. Because of the geometry of the guard it needs fitted with front guards, it is essential, in the authors' 

a horizontal bumper force of about 70 kN to start opinion, that they should be capable of absorbing 

collapsing which rises to about 130 kN at the end of energy. 

its stroke3. The ground clearance of the guard in all Because some designs of energy absorbing guards 

tests was 300 mm. 

may be speed sensitive (for example, hydraulic systems) 

the test should be dynamic and at a speed 

representative of severe but survivable impacts. The 

test procedure developed is similar to that proposed in 

ffi 

Reference 2, usirrg a trolley impactor. Threc differcnt 

types of energy absorbing guard have been used in the 

test s. 

&&r 

e m " 

E 

ff*#' - 

G 

ffii,, 

Figure 1. Invgrtube tront underrun guard 

Moving Car to Moving Truck ImPact 

Tests 

Method of testing 

Two tests were performed with both the truck and 

the car moving at the same speed. A steel cable was 

attached to the rear of the lorry and, via a pulley 

fixed to the ground behind il, Lo the front of the car' 

Thus the car was pulled towards the truck as the truck 

moved forward, resulting in both vehicles moving 

towards each other at the same speed. The truck was 

707

Results from the test with no truck guard 

fitted 

In this test the height of the standard bumper of the 

truck was 580 mm from the ground. The impact 

closing speed was 64 km/h. The deceleration traces 

for each 'B'-pillar were found to be similar as regards 

general shape and maximum values. One of the two 

traces is shown in Figure 2, together with a note of 

important events at the times they happened. The 

latter were obtained by matching the deceleration-time 

trace with the events observed on the high speed film. 

The peak deceleration of the car was found to be 38 

g. This peak occurred when the standard bumper of 

the lorry made contact with the base of the 'A'-posts 

of the car (see Figures 2 and 4). The maximum seat 

belt load was found to be 2.47 kN. However this low 

value was due to the truck bumper reaching the 

windscreen of the car and penetrating the occupant 

compartment, thus causing contact between the upper 

parts of the dummies (including their heads) and the 

truck bumper and reducing forward movement and 

belt forces. In such a case, in spite of the restraining 

708 


driven by an experienced stunt driver: the car contained 

a dummy driver and passenger. 

The intended position of impact of the car was into 

the centre of the truck guard and at right angles to it. 

To facilitate this, the left side front and rear wheels of 

the car were placed in a guide rail fixed to the 

ground. Just before impact with the truck the car left 

the guide rail and the towing cable was disconnected 

by means of a bomb release. The car was therefore 

free running when impact took place. Both the car 

conrscr w,rh 

I I tavu!;on of ruck bumos, 

and the truck speeds were measured just before 

impact. 

In these tests the lorry had an all up mass of 5300 

80 100 1?0 140 160 18d 200 

kg, and the cars used were British Leyland Minis, 

each of total mass 750 kg. Anthropometric dummies 

were placed in the car front seat positions and were 

restrained by three point inertia reel seat belts. No 

Figure 2. Car 

accelerometers were placed in the dummies. However 

one was placed at the base of each of the two 

'B'-pillars 

of the car. The seat belt loads across the 

chests of the dummies were also measured. The 

deceleration of the lorry chassis was measured and 

each collision was filmed with high speed cameras 

running at approximately 400 lrames per second. 

The intended speed was 3? km/h per vehicle to give 

a closing speed of 64 km/h which had been used in 

the moving car to stationary truck tests. 

The first impact was carried out with no underrun 

guard fitted. The test was then repeated with the truck 

fitted with a Tl Tube Products invertube energy 

absorbing guard described earlier in the introduction 

and shown in Figure l. 

'B' plllar deceleratlon pulse for test wlth 

no guard fltted 

action of the seat belts, there would be little hope of 

saving the front seat occupants. 

The length of thc vehiclc crushed by the truck was 

1.32 m (43 per cent of the total length of the Mini). 

The damage to the car may be seen in Figure 4. 

Results from the test with a truck guard 

fitted. 

This te$t was undertaken with a TI Tube Products 

invertube front underrun guard fitted to the truck. 

The lowest edge of the bumper was 300 mm from the 

ground as can be seen from Figure l. 

The impact closing speed in this case was 62 km/h. 

One of the 'B'-pillar deceleration traces for the car is 

shown in Figure 3. The peak deceleration was found 

to be 33 g. From the high specd film analysis this 

peak occurred when the bumper made contact with 

the engine block as noted on Figure 3. 

The peak belt load recorded was 5.56 kN which is 

considered to be a survivable belt load. 

The maximum amount of crush was 0.65 m (21 per 

cent of the original length ol'the Mini) and this gave 

: ? 0 

E 

GuB.d I'rrv corraoles 

m 40 6d 80 too 

Time (Elec) 

I ?0 140 160 1BD ?00 

Flgure 3. Car 'B' plllar deceleration pulse for test with 

guard fitted

, 

* ',t 

F 

& 

Figure 4. lmpact t€sts without truck guard fltted. Both 

vehlcles moving before imPact 

no intrusion into the occupant compartment by the 

truck front. There was however slight deformation of 

the car footwell and at the centre of the car facia. 

The truck guard invertubes collapsed fully in this 

test. Also, bending of the bumper beam occurred 

which suggests that at least 20 kJ of energy was 

absolbed by the guard system. The results from this 

impact can be seen in Figure 5. 

Comparison With Moving Car to 

Stationary Truck Tests. 

Provided the two vehicles remain togcther after 

impact, the decelerations and resulting damage to the 

car should be dependent only on the closing speed of 

the vehicles. It should not matter whether the truck is 

moving before impact. 

Comparisons of energy changes 

The energy changes that occurred in the four 

impact tests, two with both vehicles moving and two 

Flgure 5. lmpact te$t with truck Euard fltted. Both 

vehlcles movitrg before imPact 


with the truck stationary, are shown in Table 2. 

Energy change here is the difference between the total 

energy of the vehicles just before impact and that 

remaining immediately after impact is completed. The 

theoretical speeds after impact have been calculated 

using conservatigrt of momentum theory. In general, 

there is good agreement between theory and practice 

for these speeds after impact. 

There is also generally good agreement between the 

energy changes that occurred in the impacts where 

both vehicles were moving before impact and in the 

impacts where the truck was initially stationary. The 

exception is the test without a guard fitted and truck 

stationary, where the initial speed of the car was 

greater than intended. 

Figures 6 and 7 show the two impacts, one without 

and one with the guard fitted, for the stationary truck 

tests. These figures are directly comparable with 

Figures 4 and 5 and generally show very similar 

results from the two methods of testing. 

Comparison of impacts where no truck 

guflrd was fitted 

ln comparing the two types of testing procedure in 

detail, the rnore important dif'l'erences occur in the 

pair of tests where the truck guard is not fitted. 

Firstly, the amount of car crush in the stationary 

truck test was l.l7 m, while that in the moving truck 

test was 1.30 m. lt can be seen from these results that 

there is a substantial difference between the two, 

Table 2. Vehlcle speeds and energy changes in impacts. 

Comparison of actual and theoretical 

r€Bults 

Impact teat 

SpEedF beFdre 

i.modc t m/ s 

Speed after 

impact m/a 

loth vehicles moving IrEk C6r Trrck Car 

lr thout 

luerd 

{ith 

I Uaro 

Irrk 

trt hout 

luard 

{ith 

Ju6rd 

Act ual 

Ieet 

Ch6nqe in 

kinetic energ, 

8,9 + 8.9 7.0 - 7,0 9r 

Theory. 8.9 + 8.9 6.7 - 6,7 104 

Act uEI 

test 

gtationery 

8,6 + 8.6 6.6 - 6,6 92 

h€ofy 8.6 + 8,6 6,5 - 6.5 i1 

Ac tual 

lest 

0 +20,1 + 2.4 + 2.4 117+1 

Theory 0 +20.1 + 2.5 + 2,5 115r+ 

Act ual 

test 

0 +1 7,8 + 2.? + ?.7 97 

Iheory 0 +1 7,8 2.2 + 2.2 104 

t Positive speed is in the oriqinsl direction of motion 

of the cer. 

r+tonpacatively larger enerqy ghanqe de tD qreetet cloeing apeed. 

KJ 

709

although both amounts of crush mean that the 

intrusion into the car has extended to well behind the 

'A'-posts. 

The reason for the greater crush in the moving 

truck test can be seen in the high speed film' During 

the impact the truck bumper follows a horizontal 

path, contact occurs between the top of thc engine 

and the bumper, causing the engine and subframe to 

twist so that the top of the engine pivots towards the 

rear of the car. This causes the car to sag downwards 

until the bottom of the 'A'-posts and sills touch the 

ground. At this point the truck meets a resistance and 

cannot crush downwards any more. However, after 

the main impact the truck still has forward momen' 

tum and therefore overrides the structure of the car' 

thus driving the front of the truck up and into the 

passenger compartment of the car. 

This 'jacking' effect was observed in investigations 

of road accidents and is described in Reference 4' 

Seat belt loads are not available for the stationary 

truck test but as in the test with both vehicles tnoving, 

contact was made betwcen the dummies and the front 

of the truck. ln such circumstances, as stated before, 

seat belt loads are largely irrelevant for survival of the 

car occupants. 

The peak deceleration in the stationary truck test 

was 45 g for the car compared with 38 g obtained 

when both vehicles were moving. Bearing in mind that 

the closing speed of the former test was higher, these 

values are in good agreelnent. 

The pitching motion of the car during the impacts, 

as observed in the film, varied slightly between the 

two test$. With the truck stationary the only pitching 

of the Mini is as the front gets wedged underneath the 

truck. The same situation occurs whcn both vehicles 

are moving. However in this case the truck appears to 

have a greater vertical displacemerrt (that is riding 

Figure 6. lmpact test wlthout truck guerd fltted' 

stationary before impact 

710 


over), and as has already been mentioned the amount 

of crush on the car was greater' 

Comparison of imprcts where a truck guard 

was fitted 

With the underrun guard fitted the amount of crush 

of the car in the stationary truck test was 0.60 m, 

while that with both vehicles moving was 0.65 m, 

giving close agreement for the two test methods. 

A maximum seat belt load of 5.10 kN was measured 

in the stationary truck test compared with 5.56 

kN when both vehicles were moving. Both these loads 

are survivable and are meaningful results since there 

was no contact between the dummies and the truck 

structure as occurred in the tests with no underrun 

guard fitted. 

A peak deceleration of 40 g was obtained in the 

stationary truck test compared with 33 g with both 

vehicles moving. The small diff'erences in speeds of 

impact and amount of average crush of the car would 

account for some of the variation in the peak decelerations. 

but not all. 

There were again some differences in the pitching 

motion of the car during and immediately after 

impact but the effect on recorded decelerations would 

be small. From the high speed film it can be seen that 

with the underrun guard fitted and the truck station' 

ary then during the impact the car itself remains 

almost horizontal thus giving true deceleration values. 

However, immediately after the impact the truck is 

moved rearwards by 2 m, and the car itself takes up 

severe pitching motion with the rear wheels being 

lifted approximately 0.5 m off the ground. 

When both vehicles are moving it can be seen that 

during the impact, the car pitches more than with the 

stationary truck. However after the main impact the 

car appears to start to pitch, as in the stationary truck 

rffiffii 

lmpact test with truck guard fitted. Truck 

stalionary before imPact

ca$e, but the car then becomes trapped between the 

underrun bumper and the standard bumper and thus 

remains reasonably horizontal. 

Summary of comparisons 

Table 3 is a sumrnary of closing speedsn the 

amounts of car crush, peak decelerations at the car 

'B' pillar and seat belt loads in the four impact test$. 

The conclusions to be drawn from these tests are as 

follows: 

l. The tests suggest that a truck front underrun 

guard with a ground clearance of 300 mm and with a 

capability of absorbing about 20 kJ of energy provides 

protection from fatal or severe injuries to seat 

belted occupants in a European small car at closing 

specds up to about 65 km/h. 

2. The energy change in an impact between a truck 

and a car is largely dependent only ort closing speed, 

irrespective of whether or not the truck is moving 

before impact. 

3. When no truck guard is fitted, the moving truck 

to moving car impact test gives a more realistic 

indication of the dangers of underrun, particularly 

regarding the degree of car crush, 

4. When a truck guard is fitted there is little 

difference in the resulting car crush, decelerations or 

dummy seat belt loads whichever method of test is 

used. Because the test procedure with the truck 

stationary before impact is much simpler and cheaper 

to perform it is recommended as a satisfactory 

method of developing truck front underrun guards 

and as a means to demonstrate their benefits in use" 

Trolley Test to Measure the Energy 

Absorption of Front Underrun Guards 

It ha$ been statcd earlier in the paper that in the 

opinion of the authors it is essential that f ront 

underrun guards should absorb energy in order to 

raise the maximum survivable closing speed for car 

occupants. The tests carried out using one type of 

truck guard have shown that it is feasible to incorporate 

at least 20 kJ of controlled energy absorption 

into the guard. 

Table 5. $ummery of illnl car perlormence In lmpact 

tests. 

loth 

,!ihic I eg 

rovrnil 

r uck 

tbtior'cry, 

Impect test 

Cl dsi r1g 

spcPd 

n/s 

Pe ak 

dece I eratjdn 

a 


Car 

cI dsh 

Peak 

seat belt 

I oad 

KN 

Without quard t 7,8 l8 l, ]0 2.41 

wiih quard 17.2 71 0,65 5.56 

wrtnout guard 20,1 45 1.17 Not 

avai I Ebl ( 

wrth qu8rd | 7.8 40 0.60 5. 10 

This section of the paper describes the development 

of a dynamic test procedure to measure the energy 

absorbed by lront underrun guards. To encourage 

future guards to have more energy absorption, an 

input energy change of 25 kJ has been used. 

The aims of the test are to monitor the force levels 

and horizontal stroke of the guard to ensure that the 

energy absorption is used satisfactorily in impacts 

between small cars and trucks. 

The test procedure 

The mobile impactor or trolley is shown in Figure 

8. It is basically a 6 mm thick piece of square section 

steel tube which measures 200 mm by 200 mm and is 

approximately 2 m long. It runs ou wheels and has 

extra ma$s bolted to it to bring its total mass up to 

250 kg. The impact face is a l0 mm thick piece of 

steel which is 200 mm high by 400 mm wide and has a 

50 mm thick piece of plywood bolted to it. This in 

effect forms a non-deformable impactor. The impact 

face on an earlier version of the trolley was 200 mm 

square, to represent a car engine block. However, in a 

trial impact into the centre of an underrun bumper 

beam, it caused considerably more bending of the 

beam of the guard compared with that occurring in a 

similar impact with a car. 

The underrun guard to be tested was mounted, in 

accordance with the manufacturer's instructions, to 

the rear of a truck of similar mass to the one used for 

the car impact tests. Thus any limitations in the 

effectiveness of rhe guard by the truck itself should be 

demonstrated. 

The trolley was propelled into the guard at a speed 

of about 50 km/h to give the required energy change 

of 25 kJ. The method of doing this was similar to 

that described earlier for the moving car to moving 

truck tests except that the trolley was towed up to 

W, 

Flgure 8- Tesi trolley 

$ 

{ rfl 

lll

speed by a powerful car using a towing cable which 

passed beneath the truck. Points of impact were into 

the centre of the guard bumper beam and also in line 

with one of the drop arms. It was arranged for the$e 

tests that the trolley body centre line was at the same 

height from the ground as the middle of the bumper 

beam. As in the case of the car tests, the trolley was 

free running at impact. 

The speed of the trolley was measured just before 

impact and transducer$ were fitted, one each side of 

the body of the trolley, to measure deceleration. The 

tests were also filmed by high speed cine camcra$. 

The underrun guards tested 

Three commercially available guards, normally fitted 

to trucks to satisfy British requirements for rear 

underrun protection, were tested. Each has built-in 

energy absorption working on different principles. 

They all employ similar structures which consist of a 

cross-beam welded to two vertical drop arms. The 

drop arms are pivoted from brackets which are 

attached to the truck's chassis. The cross-beam or 

bumper bar is a rectanglar hollow steel sectiotr 150 

mm by 75 mm with 6 mm wall thickness. The energy 

absorbing mechanisms used in each guard are described 

below. 

The TI Tube Products 'Rearguard', shown again in 

Figure 9(a), has been described briefly in the introduction 

to this paper but in more detail, it works as 

follows: 

It uses two invertube cartridges 450 mm long and 

capable of being compressed by 200 mm. These join 

the lower ends of the drop arms to brackets mounted 

on the truck chassis approximately 300 mm forward 

of the drop arm pivot brackets. In the event of an 

impact the drop arms swing forwards, causing the 

invertube cartridges to operate when the compressive 

force in either reaches about 50 kN. The function of 

these cartridges is to absorb the energy of an impact 

by turning a steel tube 'inside out'. The system may 

be tuned by changing the dimensions of the invertube 

cartridge to enable it to absorb the required amount 

of energy. In the form tested the pair of cartridges 

could absorb about 20 kJ of energy. Additional 

energy may be absorbed by bending of the bumper 

beam in severe impacts. 

In the Quinton Hazell 'Underider' guard shown in 

Figure 9(b) the cartridges used operate on a different 

principle. Instead of using invertubes, the cartridges 

contain a special hydraulic oil. When the cartridges 

stroke in an impact, the oil is forced through a 

profiled slot in the cylinder wall and a resisting force 

is produced which is dependent on the speed of 

impact. Thus its energy absorption increases with 

increasing speeds. Externally the cartridges have 

strong coil springs which are designed to return the 

712 


cartridges (and thus the drop arms and the crossbeam), 

back to their original state after an impact. 

Tests on this guard rvere carried out with two pairs of 

cartridges having diftering internal profiled slots. 

Finally, the 'CURB' underrun guard, made by 

Lostock Hall Fabrications Ltd, is shown in Figure 

9(c). It does not use impact absorbing cartridges as in 

the two previously described systems. Instead, at the 

top of each drop arm there is a butyl rubber block 

placed between the arm and the bracket that it pivots 

from. In an impact the drop arms pivot and meet 

resistance from the rubber. Because of the type of 

rubber used, energy is absorbed. After impact the 

guard is designed to return to its original position. 

Trolley test result$ 

A total of eight impact tests were performed. 

Complete results are not available for two of the 

guards. Filming was not possible due to weather 

conditions for the centre impact into the invertube 

guard and there was insufficient time to carry out the 

CURB centre impact. 

The full results for the remaining six tests are given 

in Table 4. The force during impact was derived from 

the mass of the trolley times its deceleration. Trolley 

displacement relative to the truck during impact was 

measured from the film and thc total energy absorbed 

by the guards was determined from the area beneath 

the force displacement curves. 

Overall, the energy absorbed by the guards increased 

with increase in maximum displacement. The 

largest displacement occurred in test 86, into the drop 

arm of one of the hydraulic guards but was caused 

mainly by a partial tailure of one of the mounting 

brackets. 

The peak forces varied with the type of guard. The 

Tube Products invertube guard and the CURB produced 

the largest forces. However, in the impacts into 

the Quinton Hazell guards with modified hydraulic 

units, the forces irrcreased nearly to the same level. 

It is only in the tests with the hydraulic guards that 

a full comparison may be made between impacts in 

line with a drop arm and into the centre of the 

bumper beam. In both pairs of tests (onc pair with 

modified hydraulic units) the force levels in the centre 

impacts were about 15 per cent greater. This would 

indicate that even in asymmetric impacts, work is 

being done by both energy absorbing cartridges. This 

was also found in the impact in line with the drop 

arm of the invertube guard, where the stroke of the 

cartridge at the drop arm away from the trolley 

impact point was at least half that of the other unit, 

showing that it had done work. 

In both the centre impacts carried out the bumper 

beam deformed, thus absorbing some energy. The 

amount of bending is less than that produced in the


impact between the small car and the invertube guard impacting face would appear to be a satisfactory 

but bearing in mind the higher impact speed and comprotnise. Virtually no beam bending occurred in 

energy in the car irnpact, the dimensions of the trolley any of thc impacts in line with the drop arms' 

rTT ,4 x 

itff$ 

wdi 

4*,fi\ 

t t 

'!*rpll 

t " " 

P--' 

h 

I' ft 

f," 

s 

fi 

; l 

,W 

d 

,i 

t 

ilt 

J 

t 

(a) Invertube energy absorber 

?Hydraullc energy absorber 

#l 

'Try 

fl 

r ilF 

il= 

-==! 

n m 

W-**.L ffi 

(c) Rubber energy absorber 

Figure 9. Three types ol 

underrun guard before and after lmpact

Table 4. Trolley test results 

Test 

rhb6r 

Point 

df 

impact 

BT Drop 

Erm 

B4 Drop 

arm 

T ype 

of 

guaro 

lubE 

Producte 

InvsrtL66 

Quinton 

HazeII 

Hydrculic 

B5 CGhtr€ Qui rrtqn 

HEzel I 

Hydrsulic 

86* Drop 

srm 

Ouinton 

Ha4l I 

HydrEuI ic 

87* Centre Quinton 

Hazol I 

Hydraul ic 

B8 Drop 

Erm 

CURB 

Rubber 

Speed 

of 

ifrpact 

n/a 

Erergy 

ebeorbed 

by guatd 

KJ 

Haximo 

Fdfce 

duri nq 

ImpHct 

KN 

Haximum 

dlBpI dcemen 

of front of 

tfolley 

ft 

+* 

It.7 10.9 121 0.19 

14.0 16.8 & 0.40 

14,7 18. o 74 D,47 

14,8 19. I 90 0.5J 

t4,2 t 6.2 Iil 0,28 

I t,6 14. 

] 121 0,29 

Different profiled slot in hydrBUIic units cmpered with thoee useo 

in tasts 84 and 85. 

* Eelativs to truck 

The results of the impacts in line with the drop 

arms, for each type of guard, may be seen in Figures 

9(a), (b) and (c). Their force displacement curves are 

shown in Figure 10. The shapes of the curves are 

different for each guard. The force in the invertube 

guard increases with stroke up to a maximum of just 

over 120 kN to give a total energy absorption of 

about ll kJ. The hydraulic guard force rises to a 

lower maximum of just over 60 kN but maintains that 

level of force over considerably longer stroke to 

absorb about l7 kJ of energy. In the CURB guard 

result, the force level is quite Iow oyer the initial O.lj 

m of stroke. The analysis of the film suggests that 

these low forces occur while the rubber insert is 

compressing. The force then rises rapiclly as the drop 

arms deform and the stroke increases to 0.Zg m. The 

total energy absorbed was about 14 kJ but most of 

z 

Figure 10. Dynamic force-deflection 

cune$ for three 

types of energy absorbing underrun guard 

714 

110 

s 

80 

30 

0 

I 

,.,",,,". .,,..,, ,'\ l*/4- 

I 

\J 

i 

i \ i 

, 

0.10 0.20 0 30 

Ilrlli.rlon m 

"'"""'"'"'sY 'h$rhc' 

{BEr 


Hynraurrc 

€noTy rbhrbr. {84) 

this was dissipated in drop arm bending (see also 

Figure 9(c)). 

It is interesting that in the six tests, none of the 

guards absorbed more than about three-quarters of 

the trolley energy change. The lowest absorption was 

obtained with the invertube guard when around half 

was absorbed. In this test neither of the invertube 

cartridges stroked fully. However, good agreement 

was obtained between the energy absorption calcu- 

Iated from the linear compression of the cartridges 

and that measured from the force displacement curve. 

It does, therefore, appear to be a genuine result. 

Detailed analysis of the deceleration traces and high 

speed film did not yield any obvious parhs for the 

remaining energy. Trolley rebound velocities were very 

low in all cases and truck movements during impact 

were also negligible. It is possible that energy was 

absorbed by the temporary bending of various compo_ 

nents of the truck such as chassis members and the 

cab structure and in wind up of the vehicle suspen_ 

sion. 

Implicntions for future legislation 

If underrun protection at the front of trucks 

becomes mandatory in the future, it is suggested that 

legislative requirements for two aspects of its perfor_ 

mance would be necessary. The first should be concerned 

with its energy absorbing properties and the 

second with its ultimate strength. 

Energy absorption. Corrsidering first the energy ab_ 

sorption requirements, a dynamic test should be 

specified similar to the trolley test described in this 

paper. The test speed should be at least 50 km,zh and 

it is suggested that the energy input should be at leasr 

25 kJ. Limits should be placed on the force levels and 

maximum stroke during the impact. As stated earlier 

the force level should lie within a range which will 

allow the frontal structure of a small car to crush. In 

the car to truck impacts described in this paper 

maximum forces of between 240 kN and 300 kN (33 

to 40 g) were obtained when the trucks were fitted 

with one type of guard. 

Relating rhese values to a 2510 kg impactor would 

suggest that the trolley or impactor deceleration 

should be limited ro a maximum of 100 g (equivalent 

to a force of 245 kN). It may also be advisable to 

have a lower limit, say 50 g, in the initial stages of 

collapse of the guard, for example in the first 100 

mm. This would extend the benefits of the energy 

absorption properties of the guards to lower energy 

impacts between cars and trucks. 

It i$ suggested rhar rhe maximum $troke of the 

guard in the proposed t€st should not exceed 400 mm. 

This should prevent the cab structure of the truck 

from reaching the passenger compartment of the car 

after allowing for the crush of the bonnet of a small

car. The only way of increasing the allowable stroke 

would be by permitting the guard to project forward 

of the front of the truck. 

Test impacts should be carried out both at the 

centre of the beam and in line with one of the guard 

drop arms. The latter test should ensure that the 

torsional strength of the bumper bar and its joints 

with the drop arms are sufficient to utilise mosr of rhe 

guard's total energy absorbing capability even in an 

offset collision. It may prove necessary to specify a 

third impact position, close to the outer end of the 

guard, to prevent weak bumper beam ends being 

used. These can allow the truck structure to $trike the 

car passenger conparfment in collisions close to the 

lateral extremities of the truck. The required guard 

strength in this area is not easy to specify. If the 

guard is too stiff it may tear into the footwell of a 

weak car. One test carried out at TRRL with a car 

impacting into the outer end of an invertube underrun 

guard produced reasonably satisfactory results. Although 

the beam end bent to some extent, and 

therefore absorbed some energy, it did not allow the 

upper structure of the truck to strike the car passenger 

compartmenf. Further work is needed in this area to 

finalise a requirement. 

The top edge of the impactor face should be no 

more than 400 mm fronr the ground to ensure that the 

guard, when fitted to the truck, is low enough to be 

effective. 

It may be questionable whether such a test procedure 

should be carried out with the underrun protection 

mounted on a truck because of possible damage 

to the vehicle. It does however have the important 

advantage that it is representative of how the guard 

will be used in practice on the road. For underrun 

protection which forms an integral part of the truck 

cab structure, it may be the only way to test it. 

An alternative test method would be to mount the 

guard via a rigid subframe to a large concrete block 

but further research would be needed to compare the 

results from the two methods of test. 

Ultimate strength. The guard or protection must be 

ultimately strong enough so that when it is struck by a 

large car it does not break off and allow the car to 

underrun the truck structure. It is obviously not 

justifiable to make the structure strong enough to 

survive a large car impact at very high speed since it is 

unlikely under such circumstances that even seat 

belted car occupants would $urvive. 

In an earlier 64 km/h test impact at TRRL between 

a heavy car (a 200 series Volvo) and the front of a 

truck fitted with a Tube Products underrun guard, the 

guard survived the impact. Since this guard meets the 

European Community Directive EEC 79/490 (which 

stipulates maximum forces and deformations for rear 

underrun protection) in order to satisfy the British 


Construction and Use Regulation 49, it is suggested 

that this requirement would form an adequate criterion 

for ultimate strength for front underrun guards. 

Testing other types of front underrun protection. In 

this paper the only type of protection that has been 

considered in depth has been the 'add-on' type of 

guard. It may, in the future, be possible to provide 

front underrun protection as an integral part of the 

cab structure or as an extension to it, in the same way 

that energy absorption is provided in a car frontal 

structure. This could prove to be both lighter and 

cheaper. 

The testing of this type of protection for legislative 

purposes may provide some problems. It would almost 

certainly have to be tested 'on the truck'. 

Whether a trolley test would be satisfactory for 

measuring energy absorption performance is not easy 

to say without further research. Ultimately the true 

test of its effcctiveness would be demonstrated only 

by impacts using a car or car sized mobile barrier 

having a deformable face. 

Conclusions 

Car impact tests to demonstrate the 

effectiveness of truck front underrun guards 

l. A truck front underrun guard with a ground 

clearance of 300 mm and capable of absorbing 20 

kJ of energy can provide protection from fatal or 

seyere injuries to seat belted occupants in a 

European small car at closing speeds up to about 

65 km/h. (Equivalent, in this case, to a car 

velocity change of 57 km/h). 

2. A test procedure, involving a small car impacting 

a truck fitted with a front underrun guard where 

both vehicles were moving towards each other 

before impact, gave similar results to those of an 

alternative procedure in which a similar car was 

impacted into a stationary truck at the same 

closing speed. 

3. The stationary truck test is therefore a satisfactory 

method of developing front underrun guards 

and demonstrating their effectiveness. 

4. The test without a front underrun guard fitted 

showed the possibility of additional crush of the 

car occurring when the truck has initial speed. 

This can be caused by the truck 'climbing over' 

the car after the basic impact is completed. 

Trolley test to demonstrate the effectiveness 

of the energy absorption of front underrun 

guards or other underrun impact protection 

A series of impact tests have been carried out at 50 

km,/h bctween a rigid faced trolley impactor of mass 

250 kg and three different types of energy absorbing 

underrun guards fitted to a truck. The results suggest 

that the following procedure could form the basis for 

715

a legislative test to determine whether a front underrun 

guard has adequate energy absorption. The procedure 

might also be suitable for testing alternative 

forms of protection such as integral low front structures 

on trucks although this has not been verified. 

l. The test should consist of three impacts. Two of 

them should be at a speed of 50 km/h and with 

an energy change of about 25 kJ between a rigid 

face impactor and the underrun guard. The face 

of the impactor should be approximately 400 mm 

wide by 200 mm high and its top edge should be 

no more than 400 mm above ground level. The 

first impact should be in line with the centre of 

the bumper beam and the second in line with one 

of the drop arms. Both these impacts should be 

,' perpendicular to the guard. A third impact into 

one end of the guard has not yet been finalised. 

2, The deceleration of the impactor should be measured 

and, for the first two impacts proposed, 

should not exceed 100 g during the test. For the 

first 100 mm of guard mcvement the deceleration 

should not exceed 50 g. The total horizontal 

movement of the guard during impact should not 

exceed 400 mm. 

3. To en.sure sufficient ultimate strength, the front 

guard should satisfy European Community Directive 

EEC 19/490 (which is the Direc-tive normally 

used to stipulat€ maximum forces and deformations 

for rear underrun protection.) 

4. At present, it is considered that the impact tests 

should be carried out with the guard fitted to the 

truck. 

Acknowledgements 

The work described in this paper forms part of the 

research programme of the Transport and Road 

Research Laboratory and the paper is published by 

permission of the Director. Thanks are due to TI 


Tube Products Ltd, Quinton Hazell PLC and Lostock 

Hall Fabrications Ltd who provided the underrun 

guards usecl in the trolley tests, to Ken Sheppard who 

drove the truck in the moving truck to moving car 

impact tests and to the TRRL personnel involved in 

the carrying out and recording of the test programme. 

References 

l. Riley, 8.S., and Bates, H.J. An analysis of fatal 

accidents involving heavy goods vehicles in Great 

Britain. Proceedings of the 8th International 

Technical Conference on Experimental Safety Vehicles, 

Wolfsburg 1980. NHTSA US Dept. of 

Transportation. 

2. Riley, 8.S., Penoyre, S. and Bates, H.J. protecting 

car occupants, pedestrians ancl cyclists in 

accidents involving heavy goods vehicles by using 

front underrun bumpers and sideguards. proceedings 

of lfth International Technical Conference 

on Experimental Safety Vchicles, Oxford t9g5. 

NHTSA. US Dept. of Transportation. 

3. Penoyre, S., Riley, B.S. and Page, M. Desirable 

structural features for the design of front and rear 

underrun bumpers for heavy goods vehicles. International 

Conference on Vehicle Structures - 

July 1984 at Cranfield Institute of Technology, 

England. Institution of Mechanical Engineers. 

4. Riley, 8.S,, Chinn, B.P. and Bates, H.J. An 

analysis of fatalities in heavy goods vehicle accidents. 

Department of the Environment Department 

of Transport, TRRL Reporr LRl0l3. Crowthorne 

l98l (Transport and Road Research 

Laboratory). 

Crown Copyright. Any views expressed in this paper 

are not necessarily those of the UK Department of 

Transport. Extracts from the text may be reproduced, 

except for commercial purposes, provided the source 

is acknowledged. 

The Benefits of Energy Absorbing Structures to Reduce the Aggressivity of 

Heavy Trucks in Collisions 

Ian S. Jones, 

Insurance Institute for Highway Safety, 

United States 

Abstract 

Because of the large weight differences that now 

exist between heavy trucks and cars, car occupants ar€ 

at risk of serious injury when they collide with large 

trucks. However, although there is little that can be 

716 

done to reduce this disparity in weight, it is possible 

to modify trucks so that the effects of the impact 

between a heavy truck and a car could be lessened. 

This paper estimates the effects of moctifying the 

fronts of heavy trucks to incorporate crushable structures 

with stiffness characteristics similar to the fronts 

of cars. Equations of motion are developed that show 

that equipping trucks with crushable zones would 

increase the deceleration distance available to car 

occupants 

in car-truck collisions by 40 percent and

educe the average deceleration to restrained occupants 

by a factor of l4. A method is provided that 

tran$psses this reduction in acceletation to a reduction 

in fatality risk using Fatal Accident Reporting System 

data for 1977-1985. An example is given that shows a 

cru$hable zone could reduce the likelihood of fatal 

injury to car occupants by as much as 33 percent. 


ln 1985 there were nearly 40,000 fatal accidents in 

the United States and 4,211 or ll percent of these 

involved heavy trucks (>26,000 lbs.). While attention 

has been given to improving the crashworthiness of 

cars to reduce the risk of serious or fatal injury, 

virtually nothing has been done to reduce the risk to 

other road users in collisions with large trucks. 

Seventy-nine percent of all fatal truck accidents involve 

other vehicles, and 42 percent involve a large 

truck and a passenger car (Figure l). Truck-car 

collisions are the single most common type of fatal 

truck collision and the deaths are almost always to the 

car occupants, 

Table I gives a breakdown of car occupant deaths 

in car-truck collisions by the impact area to the car 

and to the truck. Of the 2,187 car occupant deaths in 

crashes with large trucks in 1985, 29 percent occurred 

in frontal collisions, 32 percent occurred from the 

truck hitting the $ide of the car, 9 percent from the 

car hitting the side of the truck, 12 percent from the 

car hitting the rear of the truck, and 5 percent from 

the truck hitting the rear of the car. Engineering 

design improvements directed toward reducing the 

aggressiveness of trucks could reduce the severity of 

many of these crashes. For example, it has been 

Figure 1. Dlstrlbution of fatal truck* crashes by number 

of vehlcles involved-lg85 FARS data 


Table 1. Charfictfflstlcs of cer occupsnt deathe In 

car-truck crashes-l985 FARS data 

established for some time that truck rear underride 

guards can reduce fatalities by effectively preventing 

cars from underriding trucks whetr cars strike the 

rear-ends of trucks. Crash protection devices on the 

sides of trucks sirnilar to those for rear undcrride 

guards could also reduce car occupant fatalities in side 

impacts. Ways to protect car occupants in collisions 

where the front of the truck hits the front, side or 

rear of the car are less obvious, but, in theory at 

least, they can be developed, This paper concentrates 

on ways of improving crash protection for car occupants 

in collisions in which the fronts of trucks strike 

the fronts of the cars. 

Theoretical Development 

.-tr""L c"""h*;.witrr--- - 

ImfEct Point Car Occupant -Fatal,rtl- 

Truck C.l N FelcFq! 

Frqnt 

Front 

Side 

fted a 

front 

0lhEr 

Front 

S ids 

Front 

Front 

Reat 

29 

6Sq l2 

r92 9 

It has been claimed that, because of the large mass 

difference that now exists between heavy trucks and 

cars, for a giverr collision it is difficult to do anything 

that would reduce the velocity change that the car 

experiences.t Although little can be done 1o reduce the 

disparity in mass, it is possible to modify the fronts of 

trucks so that the injury potential of the impact 

between a heavy truck and a car could be lessened. 

The velocity change experienced by the car in a 

car-truck collision cannot be changedn but the distance 

and time over which it takes place can be increased by 

modifying the front of the truck. Increasing the time 

or distance of the velocity change has the effect of 

decreasing the deceleration experienced by the car 

occupants and consequently their risk of injury. This 

can be accomplished by putting a structure on the 

front of the truck with energy absorbing characteristics 

similar to those of a car. 

Consider a symmetrical head-on (or central) collision 

between two vehicles as shown in Figure 2. The 

following notation is used: 

E, (Er) = energy absorbed by vehicle I (vehicle 2); 

Kr (KJ : crush stiffness of vehicle I (vehicle 2); 

mr (mz) : mass of vehicle I (vehicle 2); 

dt (dt) = crush distance on vehicle I (vehicle 2); 

F, (F ) = force on vehicle I (vehicle 2; and 

ar (aJ = deceleration of vehicle I (vehicle 2), 

1 1 n 

280 13 

totAl 2,161 100 

q 

7t7

Figure 2. Central collision betvueen two vehicles 

During the collision the forces on each vehicle are 

equal and opposite so that Fr = Fz. If it is assumed 

that the force on the vehicle is proportional to the 

vehicle crush. then 

F1 = K1d1 : Fz = Krde. (l) 

The energy absorbed in crushing vehicle I is given by 

Er : Krd?/2' 

The energy absorbed in crushing vehicle 2 is given by 

E2 = Kzdi/z. 

Then the total energy (E) absorbed during the collision 

is given by 

E:Er+Ez 

: Ktdl/z + r-2dl/2. 

In a truck-car collision, the crush on vehicle 2 (the 

truck) is zero and the total energy absorbed is given 

by 

E : Ktd?/z. 

Alternatively, the crush on vehicle I is 

d' = rffif 

If the truck is modified so that its stiffness is equal to 

that of the car, the truck absorbs a proportion of the 

energy and the total energy absorbed is given by 

E=Ki'1/2+Kzz/z. 

where d'1 is the new crush of vehicle l. However, 

Equation (l) shows that if the vehicles have equal 

stiffness the amount of crush on each vehicle will also 

be equal such that d'r : dr so that 

E : K,(d',)2. 

Rearranging Equation (2), the crush d'1 is given by 

d', = ffi, 

However, the tgtal crush between the two vehicles is 

2d'r : 2 \iElKr. 

Note that the energy E, absorbed in the collsion, is 

the same irrespective of whether the truck is modified. 

Also the velocity change, AV, that the occupants of 

the car undergo is the same, but when the truck is 

modified the car occupants are decelerated over a 

Ionger distance so that their average deceleration is 

718 


(2) 

Tabte 2. Nurnber of car occupant end truck occupant 

tatalities in latal truck crashes by size oi car 

involved-FAH$ data 1977-1985 

Truck ft4eFnt &r ftcuptnt Clf/Tcqqk Mran Cir 

qls_fEE:f!! ___ lies Frt4rrrieg Fat+!{L_.Mass'1bsr 

1,500-r,999 17 595 35,0 r,85I 

:,000_:,499 35 91i 26-1 z,ifi 

2,500-:,999 34 I.0?? 30-1 2,721 

1,000-1,499 {9 1,60t 32.7 1,240 

3.500-1,e99 52 1,713 24.5 1,141 

4,000-{,{9S 29 156 26.1 4,211 

4,500-4,sqs t4 245 I7.5 4,750 

RcqceBsion analysis rstultt: frtrlily ratio = 4.0043 (ffi$) + {1,{3; 

r ! - 0.8. 

less. The ratio of the two deceleration distances 

2d'1ld1. Then 

2d't/dt = 2rlENt/"lZE/K, = € 

Thus, if the truck is modified so that its stiffness is 

equal to that of the car, the occupants of the car 

derive about 40 percent more ride down, i.e., the 

distance in which they are decelerated to rest is 

increased by 40 percent. To determine what effect this 

increased ride down has on the likelihood of injury 

for the car occupants, the most straightforward approach 

is to see how their average deceleration is 

reduced. Assuming a force is proportional to crush 

relationship, the average force F on the car during the 

impact is given by 

F : I 

fdt 

Jo 

rxoxzd,, 

where x is the crush distance at any time during the 

impact, and dq is the final crush distance. Then F = 

Kd,/2. 

Comparing the average force on the car in the 

unmodified truck collision (F) with that in the modified 

collision (F*) we have 

F^/F = 4ntwrl-'lzn/Kt : L/E = 0.7. (3) 

Table 3. Number of cer occupant and truc* occupant 

fatalities In fetel lrontal car-truck crashes by 

size of car Involved-FARS data 1977-1985 

Eatio 

Ca. Truck kcuptnt tur OccuFn\ Cac/T.uck Mean Gr 

!CCg-1EE_-______14!-S! _ Frtrtitr+g {atalities_-Mns+-lbq. 

1,500-1,9t9 3 lEE 62.7 LE5{ 

2,000-u,499 | 328 82.0 :,?67 

i,5oo-r.sge 4 32: 80.5 2,7oj 

3,000-3,499 7 441 63,0 3,?ll 

3,500-1,999 1 32e 47,0 3,116 

f.000-4.499 6 169 26-2 4,204 

4,500-4.9T9 z 15 31.5 {.llr 

Rsgrsssion rMlytis rssultsi frtrlily Frtio = -0 0163 (firtt + lI0.li 

r + -0,82{-

However, because the mass of the vehicle is unchanged 

the ratio of modified versus unmodified 

average force also represent$ the change in the deceleration 

that the car occupants will experience providing 

they are restrained, i.e., modified deceleration = 

0.7 (original deceleration). 

Thus, equipping the fronts of trucks with a crush 

zone of stiffness equal to that of cars will reduce the 

deceleration that restrained car occupants experience 

by a factor of 1.4. Unrestrained car occupants will 

derive less benefit from the ride down. 

Estimated Reduction in Fatality Risk 

From Modified Truck f'ront-End 

Design 

If the deceleration of the occupant can be reduced 

by 1.4, injury severity will also be reduced particularly 

if the occupants are restrained. In effect, the increased 

crush space gives the occupant more distance in which 

to decelerate while the velocity change for the car 

remains the same. Another way to look at this 

reduction in deceleration is as an increase in the 

elfective mass of the car. 

This can be seen by rearranging Equation (3): 

F- : Fr.E so that F-/m = Frm Vz 

or d/ = F/m 42. 

where a' is the reduced deceleration. Hence for the 

purposes of calculating the effect of the crushable 

zone on reducing the risk of fatal injury, the reduced 

deceleration is effectively achieved by increasing the 

mass of the car by a factor of V2, i.e., the effective 

massm' : mVZ. 

There is evidence that the risk of fatality in 

car-truck collisions increases as the mass of the car 

decreases.r lf this effect can be quantified, the reduction 

in fatalities that could be achieved by increasing 

the effective mass of the car could be assessed. 

Using Fatal Accident Reporting System (FARS) 

data for years 1977 to 1985, Table 2 gives the ratio of 

car occupant to truck occupant fatalities for fatal 

accidents involving large trucks (>26,000 lbs.) 

analysed by subcategories of vehicle weight in increments 

of 500 lbs; the mean rnass of the cars involved 

is given for each car weight category. Figure 3 shows 

the fatality ratio plotted against the mean car mass 

for each weight category. The increasing fatality risk 

with decreasing car size is clearly evident, and the 

strength of the relationship is conlirmed by the 

regression analysis correlation coefficient of -0.8 

(p

injury to the truck driver. ttris would ihcrease the 

denominator in the fatality ratio and produce an 

anomalous effect of increasing fatality ratio with 

decreasing car mass. One would intuitively expect the 

effect to be small because, at closing speeds of 60 

mph, the increase in velocity change experienced by a 

40,000 lb. truck in a frontal collision with a 2,000 lb. 

car compared to 4,000 Ib. car is less than I mph. 

To examine these effects, multiple logistic regression 

procedure$ were used.o To fit the regression 

models the CATMOD procedure from the SAS Institute 

was used.s The dependent variable in each 

observation was the presence or absence of a fatal 

injury, and the independent or predictor variables 

were car mass, year of accident, and the interaction of 

car mass with year of accident. Car mass was centered 

about its mean by recoding car mass as the difference 

between the mean mass and the mass of the accident 

involved car. Regression equations were used to predict 

the log odds of occupant lhtality for the following 

situations: death of a car occupant given a fatal 

head-on car-truck accident; death of a car occupant 

given a fatal head-on car-truck accident in which the 

truck driver died; death of a truck occupant given a 

fatal head-on car-truck accident. The results are siven 

in Tables 4 through 6. 

The odds of fatality for car occupants (Table 4) 

increased significantly with decreasing car ma$s 

(p

Cs 

hErhs 

harhs 

1600 20tr 24m z8m 3?m 36m aom tffi 4ru 

tsr MsEg . tbE 

Flgure 4. Ratlo of car occupant to truck occupanl 

deaths versue car mass-frontal crashes 

the risk of fatality to car occupants in car-truck 

collisions by about one-third. Although rhe concept of 

crushable zones or front guards has received little 

atfenlion in the United States, their feasibility has 

been a$sessed in Europe. A United Kingdom study 

estimated that about 25 percent of car occupants 

killed in car-truck collisions would have survived with 

a suitably constructed front energy absorbing guard.6 

A prototype guard was constructed with a crushable 

distance or stroke of 40 cm (15.7 in.). This was based 

on the fact that 50 to 60 cm is the maximum crush 

distance available in the smallest mini-cars before 

serious intrusion starts to occur. The characteristics of 

the prototype bumper allowecl it to start to stroke at a 

load of 146 kN (approximately l0 tons). However, 

crash tests of a 1,000 kg car irrto a 5,100 kg truck at 

50 km/h (angled offset) and 64 km/h (perpendicular 

offset) showed that the bumper did not stroke fullyn 

although it proved effective in preventing underrun 

and passenger compartment intrusion. A central perpendicular 

test of a 1,550 kg car into the truck at 56 

km/h also failed to stroke the bumper fully. To 

optimize the energy absorption the load required to 

stroke the bumper was reduced to 68 kN. The 

modil'ied bumper stroked fully, in a central impact 

with a 1,fi)0 kg car at 65 km/h (40 mph). This stucty 

clearly demonstrated that protection for car occupants 

at closing speeds of 65 km/h was possible providing 

occupant restraints are used. In the absence of passive 

rcstraints or if seat belts are not worn, protection 

cannot be provided for unrestrained occupants at 

relative speeds much above 40 km/h (25 mph). 

A Cerman study that looked at the possibilities of 

reducing the consequences of car-truck accidents 

found that more than half of car-truck fatalities occur 

to car occupants involved in frontal collisions with 

trucks.T The study concluded that, although the relative 

speeds in frontal collisions were high, reducing 

the agressitivity of the front end of rhe rruck would 

also reduce the risks in truck collisions- In fact. if the 


results of the United Kingdom study (i.e., fatal 

collisions below 65 km/h are survivable with front 

energy absorbing guards) are applied to the German 

accident data, more than 40 percent of the fatal 

frontal collisions between cars and trucks would be 

affected. 

The results of all three studies lead to the conclusion 

that energy absorbing burnpers on the fronts of 

trucks could provide a substantial saving of life. Over 

2,000 accidents in the United States each year involve 

a car-truck collision in which there is fatal injury to 

the car occupant. Table I shows that in 1985, 624 ol 

these collisions were frontal impacts so that 32 percent 

or about 200 fatal accidents per year could be 

ameliorated by energy absorbing bumpers. 

Conclusions 

l. If trucks were constructed with front energy 

absorbing guards so that their energy absorptiorr 

characteristics were similar to that of cars. the 

severity of injuries to car occupants involved in 

carltruck accidents could be substantially reduced. 

2. Energy absorbing zones on the fronts of trucks 

could increase the ride down distance available to 

car occupants by 40 perc€nt and reduce their 

average deceleration by a factor of I.4. 

3. Energy absorbing zones on the fronts of trucks 

could reduce the risk of fatal injury in car-truck 

collisions by as much as 33 percent. 

Acknowledgments 

The author would like to tharrk Dr. Maria penny 

and Mr. Marvin Ciinsburg for programming thc FARS 

analysis and Dr. Paul Zador for advice with the 

statistical analysis and helpful comments regarding 

their interpretation. 

References 

l. Eicher, J,P., Robertson, H.D., and Toth, G.R. 

Large Truck Accident Causation. National Highway 

Traffic Safety Administration Technical Report 

DOT H5-806-300, July 1982. 

2. Jones, I.S. The Effect of Impact Type and 

Vehicle Velocity on Vehicle Crush Characteristics. 

Proceedings, ?lth Annual Conference, American 

Association of Automotive Medicine: 2ll-230, 

San Antonio, Texas, 1983. 

3. National Highway Traffic Safety Administration. 

Fatal Accident Reporting System 1983. A review 

of information on fhtal traffic accidents in the 

U.S. in 1983. U.S. Department of Transportation, 

National Center for Statistics and Analysis. 1983. 

4. Walker, S.H., and Duncan, D.B. Estimation of 

the probability of an event as a function of 

721

5. 

6. 

several independent variables' Biometrika. 54:.167- 

179. 1967. 

SAS Institute, Inc. 1985. S'4S Users Guide: Statistics, 

Version 5 Edition. SAS lnstitute, Inc., Cary, 

NC. 

Riley, 8.S., Penoyre, S', Bates, H'J' Protecting 

car occupants, pedestrians and cyclists in accidents 

involving heavy goods vehicles by using 

front underrun bumpers and sideguards. I1th 


International Technical Conference on Experimental 

Safety Vehicles. Oxford, July 1985. 

7. Danner, M., and Langweider, K. Results of an 

analysis of truck accidents and possibilities of 

reducing their consequences discussed on the basis 

of car-to-truck crash tests. Proceedings of 25th 

Stapp Car Crash Confe,'ertce. Warrcndale, PA: 

sAE, 198r. 

The Global Approach for Safety in the V.I.R.A.G.E.S. Project 

Plerre Soret, 

Renault Vehicules Industriels, 

France 


During the loth ESV conference (Oxford) RE- 

NAULT V.I. presented its Research Programme 

V.l.R.A.G.E.S. (Industrial Vehicle Improving Energy 

Consumption and Safety) and the test results of a first 

stage experimental vehicle (VE l0). 

This research programme directed by RENAULT 

V,I. has engaged also related industries such as 

Fruehauf and Trailer companies for the semi-trailer. 

This programme is planned from 1982 to 1988 and 

is partially aided by French Administration : 

Ministbre des TransPorts 

Agence Frangaise pour la Maltrise de I'Energie. 

V.LR.A.O.E.S. is a project with multicriteria objectives, 

which tends to satisfy new general specifications 

in all components : 

for mobilitY and energY 

for safety 

for goods transportation 

for comfort, ergonomy and live conditions 

on board 

for environmental conditions 

for industrial building. 

The aim of this project is a new global economical 

optimum in the industrial field and for the performance$. 

We present here the main elements of the second 

experimental vehicle (VE Z0) essentially on the 

SAFETY aspect. 

V.I.R.A.G.E.S. VB 20 and Primary 

Safety 

The unit has a total streamlining body and presents 

a non-aggressive outline. 

722 

The width and height conform with the new European 

regnlation but its bigger length shows the necessity 

of a better adapted regulation to satisfy the use of 

new long trailer (13.4m) and to maintain a good level 

of comfort in the cab, 

The maxi gross weight is 44 tonnes, and the unit 

conforms with the European turning circle, despite its 

important length (17.4m). F'or the tractor two possible 

solutions are in test-a 6 x 2 outline and a 6 x 4 

one. 

Handling-Roll-Over ComPortment 

The vehicle is a 6 axles unit-3 for the tractor and 3 

for the semi-trailer-with new tires and wheels with 

little diameter (wheel 19"5; height under charge, 937 

mm). 

This choice, organised with the general architecture, 

and the level controlfed hydraulic suspensions permit 

to have the gravity center 173 mm down and the roll 

axle up. 

The use of single wheel on each side of axles makes 

the real gauge larger. 

The direction control system is realised with a 

two-way hydraulic safety unit. 

In association with a very front axle, all these 

elements permit a better handling of the vehicle 

particularly to avoid the roll-over situation. 

Braking 

The energy absorption for the modulation of the 

running speed is realized by a hydro-mechanical 

retarder installed between the engine and the gear 

box. It permits to eliminate 170 KW in a permanent 

retarding phase and 300 KW during short phases. 

The braking is obtained by disc brakes with high 

pressure hydraulic control on the tractor and pneumatic 

control on the semi-trailer. There is a double 

control circuit on each axle, and an anti-slipping 

device on each wheel. 

In the 6 x 2 version an original device derived of 

the use of hydropneumatic suspensions, permit the

transient overloading of the motorised axle to increase 

the limit adherence for the starting phase. 

This micro-process piloted device acts automatically 

and uses certain components of the anti-slipping 

system. 

Yisibility-Lights 

During night the visibility of the unit is reinforced 

by a reflecting painting belt. 

At the driving place, the visibility is very improved 

because of the front and lateral windshield which 

corne down the level of the cab floor*the visibility 

distance is divided by 2 in relation of the actual cabs. 

The lateral rear visibility is realised by classical 

large mirrors. 

The Driving Place is nn important 

component of the SAFETY 

In the objective to obtain a large field of visibility 

the dash-board is dirnensioned only for the minimum 

number of instruments asked by the regulation. The 

other informations for diagnostics or vehicle management 

(fuel consumption for example) can be obtained 

on a satellite display, only when the driver asks for 

them. 

Less Splash and Spray 

The streamlining of the whole vehicle is a good 

solution to minimize splash and spray. More an 

original box betweerl each wheel collects water and 

pours it out, under the body, out of the track of the 

wheels. 

An electronical device to survey the pressure of tires 

complete this set of systems improving in a full 

manner the different factors of active safety. 

V.I.R.A.G,E.S. VE 20 and Passive 

Safety 

ln this area the specifications aim to lower seriously 

the consequences of a front-to-front shock between a 

personnal vehicle and VE 20. 

In this type of collision, the priority is to obtain a 

good compatibility with the front $tructure of the two 

vehicles, avoiding thc usual underrunning of the 

passenger car and permitting the structures of the car 

to work in the $ame manner as in a front-to-front, 

car-to-car shock. 

This is obtained by a special front structure of VE 

20. The planes of the strengths during the shock are 

compatible on the two vehicles. More, deformable 

elements of the front axle system are active between 

the bumper and the rigid body of the tractor. 

SICTION 4. TECHNICAL SESSIONS 

Shock tests at a speed of 5i km/h of a car again 

the simulated fixed front parts of VE 20 have given 

similar results as the normalized shock of the car at 

speed 50 km/h on the Eurttpean wall. 

For the lateral and rear collisions the very low and 

rigid belt gives a better protection as the actual 

devices. 

For the persons on board, a lattice structure of the 

cab, a very high position and a large inner space ( 9 

m3) aim to obtain a minimisation of the importance 

of the injuries when the cab is concerned during the 

accident. 

Safety and Use 

When the vehicle is stopped for activities related 

with goods transportations, original solutions with the 

aim of SAFETY have been built. 

The access of the cab is realised only on the right 

side of the vehicle by an inner stair-case, protected by 

sliding doors, to avoid down on the traffic sicle and 

put an end to the risks of sliding. 

Howevern these choices claim for a long cab not 

possible with the actual European regulation and the 

use of long semi,trailers. 

Tractor and semi-trailer are equipped with a sy$tem 

permitting the automatical coupling and uncoupling 

operations. This manoeuvre is initialized at the driving 

place. During th€ uncoupling operation, for example, 

down the stands, unlocking the king-pin, down the 

tractor suspension, putting the park brake, out electrical 

and pneumatical connections are automatically 

realized and the tractor is authorized to run, just only 

all necessary sequences are ended. 

Conclusion 

The experimental vehicle VE 20 is a technical 

compromise in regards the aimed multicriteria objec_ 

tives. 

It is not a commercial version, but the essential 

industrial and economical constraints have been taken 

in account to build a realistic unit. 

Despite of these constraiflt$, the vehicle presents in 

all ureas of the SAFETY, solurions permitting important 

improvements but these gains are the results of a 

global approach, that is to say a new global design of 

this type of unit in opposition of margin partial usual 

responses. 

More, these new solutions are often possible only in 

the perspective of new regulations better adapted t'or 

the goods road transportation of the years 2000.

724 

l--- 

_1.._. 

ll FRoNTAI_ 

NO RIGID 

PART 


EUROPEAN 

TIJRNING CIRCLE 

1- .11-_-!,1*1 

- s . E r * l


Side and Rear Marking of Trucks With Passive Materials 

Hans-.Ioachim Schmidt-Clausen, 

Technical University Darmstadt, 


Abstract 

Starting with an investigation about the distribution 

of the reflection factor of trucks in Europe and about 

the background luminances around the trucks during 

nighttime, the influence of different markings on the 

conspicuity is tested. Out of these, in the first step 

static experiments an optimal side and rear marking 

of trucks is developed. 


During nighttime driving, the trucks are marked not 

too conspicuous, so a lot of side- and rear-impacts 

can occur. hr down-scaled experiments special markings 

of trucks should be developed to improve the 

conspicuity. 

Figure 1. Side and rear elde of a truck In the Etreot 

L", LL, L", L": background luminances 

L,, L", Lu: background luminanceg 

Luminances of Trucks During Nighttime 

In Figure I the situation of recognition of trucks 

during nighttirne driving is plotted schematically. 

The truck is only seen in the contrast to the 

surrounding luminances Lr, Ln, L., Ls in the side 

situation, L,, Lr, Lr in the rear situation. The dotted 

area describes the luminance area neces$ary for the 

detection of a truck. 

The influence of the rear position lamps is neglected 

in these expcriments because this marking irr Europe is 

the same for all vehicles. 

In Figure 2 the test results for the threshold 

luminances for trucks seen from aside and from the 

rear are plotted. ln addition the curves for seen 

luminances in the low beam situation are shown for 

the area for the chassis (L, in Figure l) and body 

work (Lt in Figure l). 

This Figure shows that for example the body work 

can be seen from the rear at a distance of 70 m, the 

chassis at 120 m. These values are for the threshold 

case without glare etc. In the normal traffic situation 

these distances are much smaller. 

Side Marking of Trucks 

In a l:10 down scaled experiment the optimal side 

marking of trucks was investigated. ln Figure 3 and 4 

(annex) these l5 different markings are shown' 

ln an asses$ment experiment (9-step-scaline) the 

optimal lumitrance of the 15 different side markings 

was derived. The results are plotted in Figure 5. 

L/cd.m*2 

Figure 2. Threehold luminance$ of trucks seen trom 

aside, chessie, body work: mean luminances 

of these areag 

1?,5

Figure 3. Side marklng of trucke 


In this Figure the optimal luminances L for the marking 3: (Figure 3) 

markings shown in Figure 3 and 4 are plotted. The marking 6: (Figure 3) 

optimal markings are marking 7: (Figure 4) 

Flgure 4. Side marking of truckg 

726 

3, 

--_- I t--il 

1 1 r l-r l-.|J 

I --l--T - I fl-I: t-r 

r---Ln I Lr'J 

15. 

{ { { 

l t t 

---l-----T\J" 

I 

L . i4.--J Lrg


L/cd m-? 

Flgure 5. Lumlnances L lor optlmal slde marklng of 

trucks 

The worst markings are 10, ll, 12, 13, 14, the 

markings with dots and triangles, 

Rear Markings of Trucks 

ln the same test-setup the 15 different rear markings 

as shown in Figure 6 and. 7 (annex) were tested. 

nl 

L--*iJ'i 

r] 

t t 

l t 

Ir- Jl 

LJ---11 

i4 

L 

l l 

i l 

t t 

r==Tl 

=-J 

7. 

rI. '+ 1 

; 

Figure 6. Rear marking of trucks 

'l 

i 

ri 

I 

r: 

l l 

l ' l 

l-,,.^l 

L_n_l 

;- 

I: 

l 

,r--l: 

l i , i 

l i ' 

t l , 

l///R\\l ,l 

i-f-Li', 

L - - 

10. 

Figure 7. Hear marking of trucks 

The results for the rating experiment are shown in 

Figure I were again the optimal luminances L for the 

15 different rear markings are plotted. The optimal 

markings are 

marking l: (Fieure 6) 

marking l2; (Figure 7) 

marking 14: (Figure 7) 

The worst case is marking 3, the marking with dots. 

0,5 10 2D 5D 10 20 17.6 _-z 

Flgure L Lumlnances L tor optimal rear marking of 

truck8 

727

Flgure 9. Rear slde ol a truck with a conture-marklng 

with passive materlals 


Optlmal Marking of Trucks 

Out of the down scaled experiments one optimal 

marking is for example the conture-marking as shown 

in 

Figure 4: marking 7 

Figure 6: marking l. 

Improvement of Side Visibility for SafetyWhile 

Turning 

Seiichi Saitoh, 

Akitsugu Hirose, 

Nobuo Shirai, 

Isuzu Motors Limited, 

Japan 

Abstract 

In Japan, fatal accidents in which persons are 

sometirnes caught in heavy-duty trucks when the 

trucks turn left bccame a big social problem starting 

around 1976. (In Japan, the driver's $eat is usually on 

the right hand side.) The analysis result of the 

accidents showed that it will be effective in reducing 

the accidents to improve the visibility to the left and 

to prevent persons from being caught under truck rear 

wheel. To remedy the situation, the introduction of 

regulatiorrs were considered which require the improvement 

of "indirect" visibility (visibility through 

mirrors), betterment of the pedestrian protection side 

guard, addition of the direction indicator larnp at the 

middle of the vehicle siden etc.. On the other hand, 

each truck manufacturer newly in$talled an auxiliary 

side window in the left door and enlarged the front 

windshield, thus improving visibility, especially to the 

left. This report describes the introduction results of 

auxiliary side window and enlarged front windshield. 

Thanks to the effects of these revisions. combined 

728 

A rear side of a truck with this kind of marking is 

plotted in Figure 9. 

with those of regulations and others, the number of 

the fatal accidents is now approximately half of that 

for 1976. 


In Japan, fatal accidents in which persons are 

caught in heavy-duty trucks when the trucks turn left 

became a big social problem starting around 1976. 

In those days, the number of such accidents was 

not large. But the other party in the accident, such as 

pedestrians, bicycles and nrotor bicycles, was in a 

disadvantaged position. Also, in many cases, death 

accidents were tragic, such as persons being run over 

and killed. Therefore, they probably aftracted public 

attention. 

As remote causes of the accident, heavy-duty trucks 

run in the city, creating the traffic situation where 

they are mixed with pedestrians, bicycles, motor 

bicycles and other vehicles at intersections. In addition, 

many intersections are narrow, and as a result, 

the difference in turning radius between the front left 

wheel and the rear left wheel becomes large, that is, 

the rear left wheel turns sharper tharr the other party 

thinks, in left turn (Fig. 1). 

ln this paper, the authors report the history of the 

revision made to solve the problem, taking the case of 

the Isuzu vehicle as an example. The 1978 and 1986

F*lfg=*-1---.-i-;-, 

- 

model year Isuzu vehicles are shown for reference in 

Fig. 2. 

Additionally, in Japan, the driver's $eat is usually 

on the right hand side, therefore, left turn means to 

turn to the opposite side of the driver, Most of the 

heavy-duty trucks used in Japan are the cab over the 

engine type. 

lil 


At Yokohana 

Ciry 

At Kanas&ki 

City 

Flgure 1. Typlcal clty intersectlon in Japan, heavy' 

duty-truck mixed wlth pedestrient and bicycle 

on a pedestrian crossing 

l97E t'bdet 

1985 !,lodel 

Flgure 2. 1978 and 1986 model year lsueu vehicles 

rm-on1 g; -f_a:; 

,*nn" r,rn,,, ,or, | . I ,ôn, nnEEL, RIflt 

- _ j'Ery;lry:h_ -. jd',**,*'^"* "* 

XIDDTB MD LEN 6? CASES 

f r.' 

RIAR HI'[LI I,EFT 

?6 iltrs tO:l! 

I LEFT SIDE 

99II | : lo casrs 

a : l0 c^srs 

hilT ND LEtr 

I C^SIS .r. 

Flgure 3. Colllslon portions of the heavy-duty trucke 

causlng eignificant left-turn accidents 

Analysis of Accidents 

Typical significant accideuts in l9?8 by heavy-duty 

trucks (partially including buses) at left turn were 

analyzed in order to study counter-measures(l). Out 

of the total 2902 cases of accidents, 355 cases were 

analyzed. Shown below is the summary of the analysis 

results. (The significant accidents means both fatal 

and serious accidents.) 

t) As shown in Fig. 3, collision was mostly with 

the left side of the truck. Especially, the 

front and left portion accounts for 7490. 

2l As shown in Fig. 4, 83Vo of the truck drivers 

who caused significant accidents did not 

recognize the other party of the accident. 

3) As shown in Fig. 5, collisions with the 

$is:ycle and the motor bicycle are many, 

accountirtg for 8890. 

Figure 4. Recognition of the other party of the eccl' 

dent at the time of slgnlficant left'turn accldents 

by heavy-dutY trucks 

729

Figure 5. Classifications of the other party that collided 

wlth a teft turn heavy-driry riuck 

)"'i""t_ 

LEFT 

5% 

26.4 n 

n) 


BEFORE REVISION AFTER RIVISION 

TIONAL l|IiNDOI|l 

(l,vrND TYPE) 

33. I in. 

(84Omn) 

Flgure L Comparlson of the cab before and after revision 

Z\ Intrcduction of the three mirror system 

ln order to meet the revised regulation, the 

side-under-view-mirror as shown in Figure l0 

was installed. At the same time, the sideview-mirror 

and the under-view-mirror were 

enlarged. 

Others 

Besides the measures to improve the visibility to the 

left, the following two revisions were made to meet 

the requirement of the regulations, thus lessening the 

extent of injuries and reducing accidents. They are 

introduced here because they are closely connected 

with the main subject. 

l) Improvement of the pedestrian protecilon 

side guard (Article I8-2, Japanese Safery 

Regulations for Roud Vehicles) 

The side guard was changed to a larger one 

with protective ability improved to reduce 

the extent of injuries due to running over by 

the rear wheel (Fig. I l). 

2) New installation oJ the mid direction indicator 

lamp. (Article 41, Japanese Sdety Regu' 

lations for Road Vehicles) 

XILIARY 

(37s.1 sq. in. 

NEW WINDOW 

In order to securely transmit the intention of 

left turn to the two-wheeled vehicle running 

by the side of the truck, a direction indicator 

lamp was added at the middle of the vchicle 

body side (Fig. l2). 

Ttt ttnlr up to rhlch tht drlvor crn confln b), thd rlrror | 59.1 ln' (lr) 

hrtlht r ll.8 ln. (0,3)r dignclqr polq plNcod on tho urqutrd (thofln I'y thd 

shs.loJ portlon beloi) Hsr iDcrduroJ {a bcloe. 

f,l' 

{f 

I;ORWARII 

A I 

I 

Haavy rluty 

t ruck 

1 .t* 

1 .oft. 

(0'.ln) 

, i''\, 

9, 8ft . 

( 3n) 

NIiW REi;ULA] II]N 

,\N\ 

I{ 

i 6.6ft, 

I f2)r) 

_.1_ 

E.P 

, ,,,,,1\ 

\' ',; ''"\\ 

\,.\,, 

\,"''', 

\i ' \ 

'' \'\.\ \ 

Revised in 

I'larch, 1979 

Figure 9. Contents of the amendment of Japanese 

Safety Regulations for Foad Vehicles (Article 

44) "Flear-View-Mirror, etc." 

+ 

FORWARD 

,t, 

I 

Heavy duty 

truck 

731

BEFORE REVISION AFTER REVISION 

Under-view-rltffgl Sidc-view-rnirror 

,0 


Flgure 10. Comparison of the mlrror belore and after revision 

Under-view-mirror ^. . 

- 5lde-Vlew-mlfTof 

\- rsurreirLdr r ,--rl4::l- 

OLD REGIJI,ATIQIL NEW RTiOULATION 

Revised in 

March, lg79 

Figure 11. Contents ol the amendment of Japanese Safety Regulations for Road Vehicles 

"Pedestrian Protection Side Guard, etc." 

Front direction 

rl 

Revised in 

March, 1979 

Rear direction 

indicator I 

Conventional 

Figure 12. Contents of the amendment of Japanese Safety Regulations for Road Vehicles (Article 41) "Direction 

Indicator Lamp" 

732 

18.2)


BEFORE REVISION AFTER REVISION 

I ott. 

(Orn) 

r6.4ft. 

[5n) 

Figure 13. Dlrect Vlsibility diagram before and revtslon 

Effects of Visibility Improvement 

Improvement 0f static visibility 

The addition of an auxiliary side window showed a 

great effect of improving direct visibility (Fig. l3). 

Thanks to this window, when a human being came 

near the left side of the vehicle, it became possible to 

recognize him or her at a distance of approx. 3.3ft 

(lm) or morc from the vehicle. Also, for the human 

being approaching diagonally in front, the recognizable 

distance was reduced to approx. 6.6ft (2m) or 

more. (Before improvement, this distance was approx. 

6.4ft (5m) or more and approx. 9.8ft (3m) or more, 

respectively.) 

Indirect visibility approximately doubled (Fig. l4). 

Especially, the visibility to the front and left of the 

vehicle was increased very much by the new addition 

of the side-under-view-mirror. 

Effect of preventing accidents 

Actual accidents were assumed and the effect of 

preventing accidents was confirmed. Fig. 15 shows 

example patterns lbr 4 seconds before collision with 

the pedestrian, hicycle and motor bicycle. Fig. l6 

shows the relative location of the pedestrian, bicycle 

I 

Heav 

truck 

*\\N 

-16.Sft. 

[-5ttt) 

-i-- r On'the ground 

*: At a height of 39.4 in 

-: Easily visible, drivcr 

(lleight: 39.4 in. (lnl 

and motor bicycle rewritten on the visibility diagram 

with the truck placed in a fixed locafion. 

From this, the following things are known, and it 

could be said that a series of countermeasures have 

large effects of preventing accidents (Fig. l7). 

' 

I 

I 

{}'=:j 

Auxiliary side 

6ft. 

window 

'\\*-+ 

NI Heavyr 

truck'r 

Oft. 

(Onr) 

(lrn) above the ground 

slightly bent forward 

above the ground) 

1) Chances of catching higher quality visual 

information by direct view increased, 

2) The time of double visual confirmation by 

both direct and indirect view increased. 

3) The time of visual confirmation only by the 

under-view-mirror (spherical) decreased. 

4l The time of complete invisibility decreased 

almost to zero, 

Effects of the Countermeflsures Seen 

From the Statistics of Traffic 

Accidents 

Fig. t8 shows the change in the number of fatal 

traffic accidents. Although total fatal accidents level 

off through the entire investigation period, the number 

of left-furn fatal accidents has been reduced in 

recent years to approx. half of that for 1976. Like 

this, the series of countermeasures have been found 

733

BEFORE REVISION 

r*.,r-F.-l ro. +rl. 

| oft. 

[5m] | rn-r 

Unde -vlew-nlrror 


NHeavyl 

truck { 

t6.4fr. 

(5n) 

Figure 14, Indirect Vislblllty dlagram belore and after revlsion 

T 

I t 

7'l* 

_Pedestrian causht in a I 

heavy truck on-the pedesJ 

I I train cr-ossing 

Pedestrian 

' F- h Side 

AFTERREVI S ION 

| 16.4ft. 

- --a" 

Oft. 

(0nl 

t.-- 

, T- _ 

i \ 

t 

l6.4ft. 

(5n) 

t. Oft. 

t 

I (on) (om) 

Side-view- 

Ont 

: At a height 

the ground 

Numerals show the time before corlision in seconrr 

Flgure 15. Typlcal patterns of collision wlth the left- turn heavy-duty truck 

734 

,.., 

l.i 

NHeavy" 

of 39.4 in. (IrnJ above 

I 

2 

J 

4 

I!fi5to 

Collision 

with the 

motor bicycle 

start 

ing simultaneous 

1y' 

with the 

truck

CASE 1 

Pedestrian 

-"-Et 

\ l 

\l 

\ 

t 

-\, 


BEFORE REVISION -t AFTER REVISION 

CASE 2 CASE I CASE 2 

Bi cycl e 

Motor bicycle 

truck truck 

i CASE 3 

,lutotor bicycle 

Figure 16. Movement of the other party of the left-turn eccident when viewed from the heavy'duty truck 

very effective in reducing left-turn fatal accidents 

although the effecfs are combined with those of the 

improved road envirortments, tightened regulations 

and other efforts by the government and with those of 

the education of the person$ ranging from children to 

heavy-duty truck driver$ by the police. 

cAsE l 

trran 

The follesin8 shqvs Figure 16 expressed in tems of tiF 

Bafora 

t""itton 

revl sroh 

Baforc 

CASE 

r6'iston 

? 

Iti cyc lc 

revl slon 

Bcfore 

tc""on 

cAsE J 

i6tor 

,.. . Attcr 

. 

ElsFsed ti0e [sec,) 

0 l z t 

D : Rcco8nlzcd only by direct vieu 

I l Rcco8nized only by lndLrec( vleH 

N : Not seen at all 

D/I : Reco8nized by both direct and indirlct vis 

(U/H) : Reco8nized by indircct view thtough under-vie{-nirrcr 

Flgure 17. Comparlson of ease ol dlscovering the 

other party ol the accldent before and after 

revision 

In this paper, report has been made of the improvement 

of the visibility to the left in the case of Isuzu 

Motors. But similar mea$ures have been taken also by 

each of the other manufacturers of heavy-duty trucks. 

Conclusion 

Addition of an auxiliary window in the Ieft hand 

door, enlatgement of the front windshield and use of 

the three mirror system were reported as means to 

improve visibility. lt is considered that these are the 

measures near to ideal at the current level of technology. 

However, no matter what improvement is made 

3 '-' 

fotil fatsl accid€nts 

L€ft-turn fstal 

El E? 83 84 85 

(sourco: Trrfflc Strtlcticc [2]) 

Figure 18. Change in total accidents and left-turn fatal 

accidents 

735

in visibility, it will be meaningless unless drivers take 

care ald confirm safety. 

From this point, studies are under way on the 

installation of the system on the vehicle which will 

warn driver$ when there are persons around the truck. 

However, as it is difficult to disringuish between fixed 

objects on the road and persons, this system has not 

yet been put into practical use. The authors hope that 

electronics will make great progress and that it will be 

applied to realize such a system in the future. 


References 

Analysis of Heavy-Freight vehicle and rank-Truck Accidents 

736 

l. 

2. 

Traffic Safety Measures Committee of Japan 

Automobile Manufacturers Association, INC. 

"Report of Actual Conditions and Analysis of the 

Accidents by Heavy-Duty Vehicles in Turning 

Left and Right" (1980) 

Traffic Bureau, National Police Agency, Japan 

"Traffic Statistics" (1976-1985) Japan Traffic 

Safety Association 

Werner Stedtnitz, 

Herman Appel, 

Institut fi.ir Fahrzeugtechnik, 

Technische Universitiit Berlin, 


Ahstract 

It is intended to analyse heavy-freight vehicle acci_ 

dents paying special consideration to tank-truck acci_ 

dents with the help of the DAIMLER-BENZ driving 

simulator in Berlin. After the interpretation of acci_ 

dent statistics the characteristics of freight- and tank_ 

vehicle accidents were formulated by ,,characteristic 

accident patterns". On the basis of these patterns real 

accidents could be evaluated and some typical acci_ 

dents (single-vehicle-accident, frontal- and front/rear_ 

collision) were chosen for our intended profound 

simulator analysis. 

A comparison between tank-truck and heavy_freight 

vehicle accidents shows a 50go higher participation of 

tank-trucks in front,/rear collisions on motorways as 

well as in single-vehicle accidents and in fiontal 

collisions on the highways (..Bundesstra0e, 

LandstraBe"). It is assumed that ,,liquid have hitherto been scarcely considered as a factor 

possibly causing road accidents. The objective of this 

study was to investigate commercial-vehicle accidents 

with special attention directed to shifting loads, and 

with the aid of calculated and experimental real-time 

simulation in the Daimler-Benz driving simulator. The 

resulting analysis performed here is part of the project 

BASIS [German acronym for 

sloshing" 

could be one of the causal factors of such tank_truck 

accidents. The interaction between liquid sloshing and 

the accident characteristics of tank-trucks shall be 

investigated with the help of the driving simulator. A 

real-time analysis of liquid sloshing demands a simple, 

yet precise mathematical description. A mechanical 

model (pendulum-analogy and spring-mass system) tbr 

the simulation of combined transversal and longitudi_ 

nal liquid sloshing of tank trucks is prqpessd; and the 

equations of motion are presented. 


In the analysis of accidents involving commercial 

vehicles, the effects producecl by shifting cargo such 

as that encountered in partially loaded tank trucks 

,,Evaluation of Active 

Safety Devices in Simulator Work"l sponsored by the 

German Ministry for Research and Technology 

(BMFT). 

Hypothesis: "Liquid Sloshing" as a 

Possible Cause of Accidents 

The point of departure for this analysis was the 

question as to whether shifting cargo-parricularly, 

sloshing liquid loads-in commercial vehicles could 

influcnce the dynamic road behaviour in such vehicles 

(specifically, tank trucks) to such an exrenl as to 

represent a cau$e of road accidents. 

An analysis of accident $tatistics concerning tank 

trucks reveals that such vehicles are, in relative terms, 

50Vo more frequently involved in road acciclents than 

other commercial vehicles. Such increased participa_ 

tion in accidents has been observed for tail,end 

accidents on German freeways (Autobahn), as well as 

for single-vehicle accidents and head-on collisions on 

national highways and rural roacls in Germany. This 

phenomenon is emphasized in f'rgs. I and i as a 

"conspicuous 

difference" in the statistics. Assessment 

of such statistics witl of course point out characteris_ 

tics particular to tank-truck accidents; however, it 

does not suffice to allow conclusions to be drawn with 

respect to the influence exerted by the load. Conse_ 

quently, statements of various authors will, to begin, 

be cited below in initial support of the hypothesis 

stated above. 

Bauer[Z] investigated the effects of shifting fluids 

on the structural and flight stability of rockets, and

later broadened his studies to cover tank trucks. With 

respect to semitrailer tank trucks, he reported as 

follows: "A liquid loading may lead to considerable 

handling difficulties if a free liquid surface exists, 

sloshing about its fundamental natural frequency, 

thus creating forces and moments acting on the 

vehicle, which may easily exhibit a multiple of the 

inertia force detennined for the same mass if it were 

considered rigid and fixed to the container."[2, Part 

I, p. 451. In their analyses of tank-truck accidents in 

Australia, Griffiths et al. reported the following: 

"Among the factors that lead to tanker rollover are 

their hieh center of gravity (c.g.), 'soft' roll stiffness, 

and sloshing of the liquid load"[3], p.21. 

Erwin et al[.4] investigated the threshold value for 

the lateral acceleration beginning at which vehicle 

rollover can be expected. Fis. 3 throws light on the 

association between rollover threshold and the percentage 

frequency of single-vehicle rollover accidents. 

Investigation here was made of the three-axle 

tractor/two-axle semitrailer in fully loaded condition: 

i.e., without possibility of the shifting of liquid loads. 

The graphical representation is based on evaluation of 


records from the USA Bureau of Motor Carrier 

Safety (BMCS), in combination with a dynamic 

vehicle-behavior model which was used to help determine 

the rollover threshold. 

The percentage share of rollovers in single-vehicle 

accidents demonstrates a direct relationship to the 

rollover threshold, as the nonlinear plot in Fig. 3 

makes clear. According to the opinion of the authors 

involved here, the threshold value is primarily dependent 

on the height of the center of gravity. 

Even though partially loaded vehicles were not 

considered in this investigation, it is nevertheless 

obvious that a progressive, percentual rise in the 

probability of a rollover in a single-vehicle accident 

occurs together with a decrease in the rollover threshold. 

A low rollover threshold, furthermore, is occasioned 

not only by a high center of gravity-but also 

as a result of dynamic fluid tbrces. 

In investigations carried out by Strandbergf5], attention 

was called to experiments conducted by 

Abramson et al[6], in pointing out the distinct possibility 

that fluid forces in the case of resonance can 

rise to values seven times as great a$ those observed 

100 

Sinqle 

- vehirlp 

d((ident 

[i 

Vt t;,fr, 

lnf ef Se( tron 

Heod-on collisron 

ac(rdenf 

1C 

ll 

1Ww 

0 10( 

10010c 

IT 

t,' l'':. 

1,. l:i 

t-..: 

t::. 

4Vt 

l'::, i.: '1"{/, t.::, 

l:.:. 

1 ... 'lw 1',., '. 

t 

l':," 

l.' 

'lw 

IW: 

1 - 

t'. 

[:::i 

lrrl 

l> l; l> 

t- :t t: rl r: 

lh 

li 

l:. 

l...., 

fr W; d:i 

V,t lt.,. t IW 

l....r 

l. i.;tW 

l= 

W 1,,,., 

l:i:; 

l:r:l' 

Nofr0nqt hrghwoys ftg19l roods Iouniy roods 

rfl 7A 

[':':f tomme rci o I vehi c les ( CV) f7j xozt dous' ror go vEhic les 

[l;;l W (HVl 

E xo n p_&_l_q1_E_q_d1n g_g Pd ph5, I 47o of o ll h 6zdrd o us - carg 0 

occidenis on noti0nol hiqhways toke ploce os sinqle-vehicle 

dcci denis. 

0otd evdluofion by TUV ttazardous-Corgo Accident Anolysi5 /1/ 

-n 

rtt 

t4HilluOlrcHRrx 

Conspictrous dtfferences between 

pdrtitipoiion of hozdrdous-corgo 

ond conmeccroI vehicles in trqffit 

drridents, with typeE of sfreets os 

refefence f0cidr 

Fig. 2 

Sted tn t tz 

737

when the liquid load cannot shift inside its container. 

The following findings were established concerning 

the overturn risk of a tank truck from the model 

experiments conducted by Norstrcim et al[7, p. 69]; 

"With 

5090 load volume it was found that the 

increase in overturning risk compared to rigid load 

could be up to rromewhat more than two times, both 

in harmonic oscillations at frequencies low enough to 

occur in normal driving and in the doublelane change 

maneuver," 

Until now, howwer, such influences on the dynamic 

vehicle behavior of tank trucks could not be 

proved in road tests. The difficulties inherent here 

were reported by TOPAS in the following[8, p. 168]: 

"Contrary 

to all expectations, sloshirrg liquid loads in 

half-filled containers exhibit only moderatc eff'ects; in 

any case, the measured angles of roll fell below those 

values observed for the fully loaded vehicle. In a 

subjective sense, even the slightest movements of the 

liquid load-for example, the sensation detected by 

humans when the vehicle is at rest-can readily be 

detected. The effective consequence for the behavior 

of the moving vehicle is, however, relatively slight." 

The dynamic vehicle measurements serving as the 

basis for this conclusion do not, however, sufficc to 

E 

t 

@u 

frf, 

a 

- t 

,1fl 

GA 

O 

o 

E 

uE 

tl 

aro 0.s Ls! 0.a 

FOLLOVER THRESFBLD (C'S) 

ralt 

rtt 

,qrrucncitrr 

738 

tlLltroIStIt lttlttt totrLoYtl 

ttlt3EoLD rtD tollovtl tccIDIII' 

rtvotsttttr {tlsfD or rxcs 

TCCIDIIT DIIA IOI Tf,I YEITS 

\t73-rrt /rl 


Ef.vlt{ EI' AL - 

Fl-s - 3 

allow an evaluation of the influence of vehicle loading 

on the behavior of such cargo vehicles in accidents. 

This question can be elucidated only by sufficient 

analysis of the entire control system complex represented 

by the driver, his vehicle, and the accident-site 

circumstances. 

Accident analysis of heavy-freight vehicles and tank 

trucks in a closed control system is carried out with 

the aid ot the Daimler-Benz driving simulator[9,10], 

in conjunction with the Institute of Automotive Engineering 

at the Technical University of Berlin. 

In a parallel test procedure, typical and "characteristic" 

heavy-freight-vehicle and tank-truck accidents 

are-on the one hand-selected and recreated in the 

driving simulator by a large number of test drivers, 

with testing in this case performed serially. In the 

parallel procedure, on the other hand, the drivingsimulation 

model is modified to recreate the movement 

of liquids inside tank trucks. 

Characteristic Accidents 

Only the analysis ol representative accidents can 

lend hope for the reduction of future traffic acciclents 

on a broad basis. For this reason, it is necessary to 

select truly "characteri$tic 

accidents" for purposes of 

accident research on a driving simulator. 

**Characteristic 

accidents" are considered to be 

those accident$ for which the parameters describing 

the accident (e.g., type of collision, type of vehicle, 

collision velocity, etc.) represent, in their combination, 

percentually salient phenomena observed in accident 

statistics. 

Fig. 4 depicts the parallel procedure for the determination 

of such characteristic accidents. 

"Characteristic 

accident patterns" result from the 

study of road accident statistics in which salient. or 

predominating, patterns can be brought to emerge (see 

FfC. 5). With the aid of these patterns, actual accidents 

can be assessed with regard to their respective 

characteristics. 

From a comparison of the accident patterns with 

actual accidents, finally, characteristic accidents were 

able to be ascertained-from which three typical 

examples were chosen for further treatment on the 

driving simulator. These examples were as follows: a 

head-on collision of a tank truck with a pa$senger car 

on a national highway, a single-vehicle accident in 

which a tractor-trailer unit loaded with hanging loads 

of meat turns over while negotiating a curve, and a 

tail-end collision of a tank tractor-trailer unit which 

runs into a construction-site vehicle on the freeway. 

For purposes of simulating the surroundings, lhe 

accident sites were recorded by video mcan$ and were 

photographed. With the aid of road plans, the course 

of the roadway is presently being reconstructed on the 

driving simulator in the form of a digital video

Evcluaiion of 

orcrdent records 

IDEKRA, TUB/HHH, BAST) 

Desrription of occidenl 

chorocterrsiics 

-n 

rl, 

Selection of orti.dentg 

Evaluoiion of 

oqcident slslistics 

Formuts,tion of 

" 

chorac lerisfrc 

occideni potterns" 

Assrgnrnenl of fhe individual 

occrdenfs to the corresponding 

chcraclerislic occidenl potierns 

Chorocteristic accidenls 

5e{ection of suitobie orrrdents 

. Heod-on collisions: tqnk truck / csr 

r Sinqte*veh. oE(-r troctor - trailer 

. Toil -end occ., tonk lruck 

Type of occidenl: 

Type of vehiqle: 

Approved totot veighfr 

Type of rood: 

Road rondilionr; 

Aqe of drivers; 

Couse of acqideni; 

Directiorr of impocl: 

Poini of impoctr 

Vehicle 

hif: 

Speed of collisionr 

.af 

Addlyrrt of Cofifr{r(rsl'vrhr(lr / 

ldnt - frvEk A

With the exception of one model based on potential 

theory (the vehiclc simulation system TDVS, plus the 

computer program SLOSH devised for liquids[13])' 

there has been until now no other spatial liquidsimulation 

program conceived for freight vehicles. 

In the following, a spatial and mechanical model is 

described which utilizes insights gained from aeronautics 

and astronautics in arriving at the most suitable 

approach for arrival at a model for the respective 

truck-tank geometry in the direction of excitation' In 

this approach, movements of the liquid in longitudinal 

and lateral directions are initially considered separately. 

They are later superimposed. 

The Pendulum Model for Tran$versal 

Oscillations 

For transversal liquid oscillations in containers with 

rounded tank walls, the pendulum analogy has proved 

to be the most effective, as can be confirmed from the 

studies conducted by Sumner[4,l5,andl6], one of the 

pioneers in work with the "Pendulum Analogy." His 

work, in addition to elaborations subsequently made 

thereon by Sayar[l7] with regard to nonlinear oscillations, 

form the basis for the model described here' 

Although Sayar and Sumner conducted their investi- 

(*-*€) o 

-E 

rtt 

PLUTD FREQT'THCY PIR.II{ETER 

SPE.ERICAI. TII{T, 

IATER.IT EXITITIOI{. 

HC CTRTI TfiD STEPEENS /19/ 


Fiq. 6 

gations on spherically shaped tanks, application of 

their findings to tank geometries featuring circular 

cross-sections is permissible, since the oscillatory behavior 

with typical tank-truck forms is sufficiently 

extensively similar to that of their containers. This 

fact has been additionally confirmed by analytical 

tests conducted by Budianski[18], as well as in model 

experiments performed by McCarty and Stephens 

[[l9]. See Figs. 6 and 7. 

The natural frequencies of the liquid, which represent 

a characteristic value for the description of liquid 

motion, demonstrate a maximum deviation from each 

other of approximately l09o for a tank filling ratio of 

h/d:0.8 (whereby h:the filline heieht and d=the 

tank diameter). 

The model for transversal liquid motion consists of 

the fixed mass-designated by mo, with a moment of 

inertia l6-which does not take part in the liquid 

movement, the mathematical pendulum with the mass 

m, which represents the fundamental oscillation, a 

damping element with the damping factor K, and a 

spring with the nonlinear characteristic C for simulation 

of nonlinear factors which arise from the curvature 

effect (a force of restoration arising, in turn, 

from the curvature of the tank). See Sayar[I7]. 

("'*#) 

al, 

h/ 2R 

ftrh \il 

o 275 t3 

o +aE t5 

O 3 l ? 

d 6 a 4 

-nrq, Ft t 

llc Cartl' rnd StdphEns 

FLUID ERBQI'E{CY P.I.RTfiETER 

EORIZOHTII CIRCIIITI* 

CYITII{DER. 

LATERIIJ EXITATIOH /19l

Fis. I defines the variables for the mathematical 

formulation of the pendulum analogy for the general 

case. The dimensions are explained as follows: 

ho 

cg 

C 

A 

1"" 

= distance from center of tank to fixed mass 

= center of gravity 

= gcometric center of the tank 

: hinge point 

= distance from the center of the tank to center of 

gravity 

q = vehicle center of rotation 

lo,o = distance from hinge point of pendulum arm to 

vehicle center of rotation 

lp = distancc from center of tank to hinge point of 

pendulun'r arm 

Lp = length of pendulum arm 

angle from vertical through which pendulum 

oscillates 

g = angular rotation about q 

h = liquid deprh 

mo = fixed mass (i.e., non-sloshing) 

Io : mom€nt of inertia of fixed,mass (non-sloshing 

mass) 

Fig. 9 depicts the external and effective forces as 

they act on the vehicle tank and the pendulum (cf. 

Sayar[ 7]). 

The equation of motion for the pendulum results 

from the law of angular momentum for the accelerated 

reference point, as follows: 

dl^/dt + mr [(r;s ",ril + (ris x fJl : E t{o 

(Eq. l) 

--]1 

rtt 

axtr.rur r0rl 

llHttludrfcHxrf, 

ltdl. tn tE.H. Aw.l 


_v/U: it, 

where: 

L; 

ris 

E M; 

* the angular ffiomentum about A 

= vector fiom the point of pendulum susp€fl. 

sion to the center of gravity of the pendulum 

mass 

= sum of the moments ahout point A 

U : the angular momentum about A 

ii;, 2; = Accelerations of the point of pendulum 

suspension 

It follows that: 

m, Lotii * mr g Ln sin p + k toTi + ca3 Ln = 

mr (Lp x^ cos p - Lp 2^ sin p) (Eq. 2) 

With sin I E I 

- pt/6 and 4 = l/6 - c,/m,g and 

the degree of damping 3 = k/Zmt on it follows that 

the equation of motion of the pendulum is as follows: 

,ii + liluoc t ,o'(e - ,tp\ : l/Lp (iia cos p 

- io sin er) (Eq. 3) 

In order to determine the forces from the liquid 

which act onto the tank, the sum of the external and 

the effective forces is found according to D'Alembert's 

Principle. See Fig. g. 

The following applies to the tank container: 

E horizontal forces (external minus effective) : 6 

- ctr3 cosp - F, - Tsin,p - kLpgco$(p + 

mo jb : 0 1Eq. 4) 

SI}IEHATII DIAERAM OF A HORIZONTAL tIRtULAN TVTITNNICIr 

CONTAINER PARTIALLY FILLEO WITII FLUIO AND ITS ANALOGOUS 

PENOULUM ANALOGY FOR LATTRAL LIOUIO SLOSHING 

Fl-tt' - a 

Stedlnilz 

741

Gr| 

rlt 

IXITITUT FOII 

tiHnltuotEcHxlx 

Prol. Or..la. H. APd 

mtxl 

mrLP,ir? 

Lp 

inrLp,i 


mt it 

nt9 

The following applies to the pendulum: 

k L p'ir 

\-Y{ 

cA3l.* 

EXTERNAL ANO EFFECTIVE FORCES FOR THE PENDULUM ANO THE TANK 

X horizontal forces (external minus effective) : 6 

T sin rp + k Le,i coslp + c p3 cos (P - m, LoPz 

sin p + m, ii, + + m, Lop cos rP : 6 (Eq' 5) 

If equations 4 and 5 are added, the total force as 

exerted on the tank is as follows: 

Fn = Inr iil + mo *o + mr Lr',p cos P - mr Lp 

i2 sinp Gq. 6) 

It must be pointed out here that, in the event that 

the vehicle is rotated about point q (see Frg. 8)' the 

accelerations x,, xA, and in will depend on the 

angular acceleration 6 (see Sumner [14 and l6])- 

The force Fr is the sum of the horizontal forces 

which are composed of the following: the forces of 

inertia of the masses, the horizontally acting bar force 

at the point of pendulum suspension (T sin .p), and 

the damping and spring force acting at the level of the 

pendulum mass. The various points at which the 

forces act must be taken into consideration for 

formulation of the moments about the center of 

rotation q. 

The approach for the vertical florce is as follows for 

the forces acting onto the tank: 

E vertical forces (external minus effective) : ff 

cp3sinp - Fv + kLe.|'sin.p - Tcosg * mo 

(zo-g)=O Gq.7) 

Fag-9 

Sled t n itz 

The following applies to the pendulum: 

E vertical forces (external minus effective) : 0 

Tcosg - kLo psing 

2,) - m, Lo iz cos ,p - 

Now, when equations 7 

following results: 

-cg3sin,p-mr(gm1 

Lorsing=6 

(Eq' 8) 

and I are added, the 

Fv = - mo(g - 2o) - mr (e - Z,) - m, Lorz 

cos .p - m, Lo

For the analysis of rectangutar tank forms, studies, 

in professional literature have featured only springmass 

simulation models, including the associated parameters. 

Fig. 10 shows such a spring-mass model 

designed in accordance with Dodge[20]. 

The model approach for the simulation of longitudinal 

sloshing in tank trucks can be maintained on a 

simpler basis, since the motion of liquids in tank 

containers with straight vertical walls is considerably 

more linear in nature than that in containers with 

curved tank walls (see Dodge[ZO, p. 205]1. 

With this model, masses of higher order can be 

neglected, owing ro their relatively slight share in the 

forces of the liquid involved. 

In the model according to Fig.10, a linear damping 

element must additionally be provided lbr simulation 

of viscosity or of bafllcs installed in rhe tank. 

The equation ol' motion of the model represented in 

Fig. ,/0 corresponds to the equation for a simple 

spring-mass-(damper) system. 

Superposition of Longitudinal and Lnteral 

Liquid Motion 

The contenrs of the individual tank chambers of a 

tank truck are each represented here by one simulated 

liquid system. These sy,stems consist of a pendulum 

model depicting lateral motion, as well as a springma$s 

system to depict Iongitudinal sloshing, in accordance 

with the descriptions given in the two sections 

above. 

1E 5IHFLE ]ITCHAHICAL 

HOOEL FOF A 

qETTAh6ULAH TANx I Ie, 


Fig, 10 

Now, if these two models are arranged perpendicularly 

to each other in a coordinate system fixed with 

respect to the tank, and if they are fixed in the 

geometric center of the tank, then they provide a 

spatial representation of the motion of the liquid. 

This representation, however, does not suffice to 

account for the mutual interaction of lateral and 

longitudinal sloshing, since the motion of the two 

models is independent of each other. In the same 

manner, assumption was made for the simulation 

systems in the two previous sections here, that the 

upper surface of the liquid is plane in the direction of 

moverlent not described in the respective model. This 

factor must be sufficiently taken into account when 

the two models are incorporated together. 

For the model approach shown in Fig. 1j, the 

attitude of the longitudinal and lateral simulation 

systems was for this reason maintained variable in the 

plane lying perpendicular to the direction of motion 

described by the respective model. 

An example will make this clearer. When a tank 

truck begins to negotiate a curve, the liquid in the 

cargo container moves in a direction lateral to the 

outside of the curve, as shown at the top of Fig. II. 

If the vehicle then decelerares, the liquid would move 

Iongitudinally toward for the direction of travelwhereby 

the initial situation provided for the longitudinal 

sloshing is represenred by the laterally banked 

upper surface of the liquid. The longitudinal 

"sloshing 

force" would then act with lateral displacement, 

and no longer in the central geometrical plane 

of the tank. Consequently, the simulation model for 

the longitudinal movement must also be laterally 

displaced, as shown in Fig. I L 

Further experiments investigating the moments involved 

will be required to determine whether this 

lateral displacement of the longitudinal liquid model, 

as represented irr Flg. 11, should be oriented to the 

lateral displacement of the center of gravity of the 

fluid. 

Now, the longitudinal displacement of the center of 

gravity determines the longitudinal shift of the lateral 

liquid model, since the lateral force of the liquidowing 

to the longitudinal sloshing of the cargo-also 

no longer acts at the middle of the tank. 

Mutual interaction between the two models, arranged 

perpendicular to each other as they are, is 

therefore involved in the manner described here. With 

displacement of thc models as described above, forces 

of inertia should not be allowed to enter the conf.emplation 

of the systems in the direction of lateral 

displacement; rather, consideration may be taken here 

of only the constantly shifting points at which forces 

exerted by the liquid are applied. 

In the case of the vertical fbrce arising from the 

pendrrlum model (Eq. 9), it must be taken into 

743

account that this force was derived from the theoretical 

model approach for lateral liquid motion' Consequently, 

further experimental confirmation ol' the 

situation here is necessary. The point at which the 

vertical force is applied is not derived from superposition 

of the longitudinal and lateral model' For this 

reason, assumption should be made of the vertical 

force acting at the center of gravity of the liquid at 

rest. 

Model Parameters 

The parameters contained in the mechanical liquid 

model presented here can bc described both analyti' 

cally as well as experimentally, as Pfeiffer[2l] and 

Sumner[4,l5,andl6l have elaborated. As fiar as the 

tank geometrical forms are involved here (rectangular 

and spherical), all of the parameters described in the 

moclel are quantitatively known and can be gleaned 

from professional literature[14,15,16,and20]. These 

parameters have been made available in non' 

dimensional form, with the result that procedures can 

be applied to geometrically similar tanks of any size' 

PENDULUH ANALOGY FOR LATERAL TIOUID SLOSHIN6 


*-l--:r*:t 

i suRFAcE 

SPRIN6-I.,IA55-ANALO6Y FOR LON6ITUOINAL LIOUIB SLOEHIN6 

aall 

rtt 

FAHfrIIUC'ECHHIX 

744 

l-,'- 

5UPERPO5ITION OF LONCITUOINAL 

AND LATFRAL LIOUID 5LOSHIN6 

-- LIOUIO 

5 UR FATE 

F ig. 11 

Itedtnitz 

Discussion 

The liquid-cargo simulation model depicted here 

represents a further development of the mechanical 

liquid models taken from aeronautic and astronautic 

engineering applications-models which were origi' 

nally conceived for calculation of the structural and 

flight stability of rockets. As a result, the question 

logically arises as to the validity of application of the 

model concept for tank vehicles. 

In the answering of this question, the study presented 

here provides closer examination of the as' 

sumptions used as basis for the liquid simulation and 

the determination of model parameter$ involved in 

tank-truck efforts. 

As has been demonstrated in the work performed 

here, assumption of the following is justified: 

. absence of friction 

. absence of cavitation 

vertical position of the upper surface of the liquid 

with respect to the pendulum. 

In accordance with investigations performed by 

Strandberg[S, Vol. lI, p. 86], friction forces in the 

liquid due to viscosity may be considered negligible in 

comparison to inertial effects. 

Cavitation-above all, at such sharp edges as encountered 

at baffles, when longitudinal sloshing occurs*can 

arise and can lead to damage in the tank' 

Since cavitation acts for only very brief periods, 

however, its effects contributing to the influence of 

the liquid load on dynamic vehicle behavior can be 

neglected[5, Vol. II, p. 88]. 

Justification for assumption of the vertical position 

of the liquid surface with respect to the pendulum axis 

was confirmed in Sumner's model experimentsll4]. In 

addition, confirmation has been provided in actual 

driving tests with a tank truck, in which the vertical 

position (referenced to the liquid surface of a physical 

pendulum mounted in a tank chamber) was indeed 

observed[22]. 

The effects of the following assumptions on the 

validity of the liquid model have not yet been able to 

be assessed; 

r absence of rotation of the liquid 

r 

. 

small excitation amplitudcs 

superposition of longitudinal and lateral sloshing 

movements. 

The assumption of freedom of rotation is of 

necessity based on the determination of parameters in 

accordance with potential theory. In reality, however, 

the liquid does in fact rotate. This phenomenon can, 

possibly, be compensated for by the superposition 

principle. 

The parameter values determined from the model 

experiments can vary according to the excitation

amplitude of the tank. Since excitation amplitudes are 

small in aeronautic and astronautic study, and since 

great amplitudes of this nature are in fact encountered 

in motor-vehicle engineering, it will be necessary in 

the future to devote particular attention to this point 

of discussiorr. 

The superposition principle proposed in this section 

must, in addition, be subjected to further study 

regarding the validity of its employment. 

Answers still open to the questions raised here can 

most effectively be provided with the aid of model 

experiments and actual driving tests, the planning of 

which has already begun. 

With regard to assessment of the hypothesis stared 

at the beginning of the paper-"Liquid Sloshing: a 

Possible Cause of Accidents?"-it would appear most 

advisable to await the results of testing perfonned on 

the Daimler-Benz driving sirDulator. 

References 

l. Jdger, P.; K. Haferkamp et al.:"Die Auswirkung 

des Sicherheitsrisikos von Lagerung und 

Transport gefiihrlicher Stoffe auf die Entwicklung 

verbesserter Transporttechnologien 

(StraBentransporr); Berichrsbdnde l-5; Kciln 1983 

2. Bauer, H.F.:"Dynamic behavior of an elastic 

separating wall in vehicle corrtainers";Part I ; 

Int. J. of Vehicle Dcsign, Vol 2, no.l, l98l; 

Page 44-77 

3. Criffiths, M.; Linklater, D.R.: "Accidents Involving 

Road Tankers with Flammable Loads"; 

Traffic Accident Research Unit, Traffic Authority 

of New South Wales, Australia, Jan '84 

4. Ervin, R.D. et al.: "Future Configuration of 

Tank Vehicles Hauling Flammable Liquids in 

Michigan" ; UM-HSRI-80-73-l; The University 

of Michigan, Decernber 1980 

5. Strandberg, L.; "Lateral Stability of Road 

Tankers"; VTl*report 138 A; Vol l,ll; 

Linkdping 1978 

6. Abramson, H.N., et al.;"Some Studies of Nonlinear 

Lateral Sloshing in Rigid Containers"; J. of 

Applied Mechanics, Dec. 1966; 5.177-784 

7. Nordstroem, 0. et al.;"Test Proeedures For The 

Evaluation of the Lateral Dynamics of Commercial 

Vehicle Combinations"; Automobil-lndustrie; 

23(1978\2; P.63-69 

L Daimler-Benz ;"ToPAS-Tankfahrzeug mit optimierten 

passiven und aktiven Sicherheitseinrichtungen"; 

Studie Sicherheitstankfahrzeug; Stuttgart, 

April 1984 

9- Drosclol,J- et al.; "The Daimler-Benz Driving 

Simulator, A Tool for Vehicle Developmcnt"; 

SAE 850334: March '85 

10. Drosdol, J. et al.;**The Daimler-Benz Driving 


Simulator; New Technologies demand New Instruments"; 

gth IAVSD Symposium on Dynamics 

of Vehicles on Roads and Tracks; Linkciping, 

Sweden, June 1985 

Stedtnitz, W., H. Appel:"Analyse charakteristischer 

Verkehrsunfdlle mit Nutzfahrzeugen"; 

l2.BMFT-Statusseminar "Kraftfahrzeugc und 

StrabBenverkehr"; Bad Ems, Sept. I986. 

Slibar, A.; Troger, H.:"Die kritischen Fahrzustrinde 

des Tank-Aufliegerzuges bei verschiedenen 

Beladungsgraden und Bertihrbedingungen im 

stanonaren Fahrbetrieb"; Bundesministerium 

fiir Bauten und Technik. 

StraBenforschung Heft 55; Wien I976 

Bohn, P.F. et al.:"Computer Simulation of The 

Effect of Cargo Shifting on Articulated Vehicles 

Performing Braking and Cornering lvlaneuvers"; 

Vol l-4, FHWA/RD-80/142, Laurel, MD., May 

l98l 

Sumner, I.E.: "Experimentally Determined Pendulum 

Analogy of Liquid Sloshing in Spherical 

and Oblatc-Spheroidal Tanks"; Nasa-TN-2737; 

N65-19919; Washington, April 1955 

Sumner, LE., et al.; "An Experimental Investi' 

gation of The Viscous Damping of Liquid Sloshing 

In Spherical Tankr;"; NASA, Technical Note 

D-1991, Cleveland, Ohio 1963 

Sumner, I.E.:"Experimental Sloshing Characteristics 

and a Mechanical Analogy of Liquid Sloshing 

in a Scale-Modul Centaur Liquid Oxygen 

Tank"; NASA TMX-999; Cleveland I964 

Sayar, B.A. et al.: "Pendulum Analogy for 

Nonlinear Fluid Oscillations in Spherical Containers"; 

Journal of Applied Mechanics, Dec 

l98l; Vol 48;5.769-712 

Budiansky, B.;"Sloshing of Liquids in Circular 

Carrals and Spherical Tanks"; J. of the Aero/ 

Space Sciences; Vol 27; March 1960, No.3; 

p. l6l-173 

McCarty et al.:"Investigation of The Natural 

Frequencies of Fluids in Spherical and Cylindrical 

Tanks"; NASA TN D-252 

Dodge, F.T.: "Analytical Representation of Lateral 

Sloshing by Equivalent Meclranical 

Models"; aus H.N. Abramson:"The Dynamic 

Behavior of Liquids in Moving Containers"; 

Chapter 6; 5.199-223; Southwest Research Inst., 

Wash. 1966 

Pfeiffer, F. ; "Linearisierte ll. 

t2. 

13. 

14. 

15. 

t6. 

t7. 

18. 

19. 

za. 

2t. 

Eigenschschwingungen 

von Treibstoff in beliebigen Behdltern"; Zeits- 

22. 

chrift frir Flugwissenschaft, l6(1968), Heft 6 

Kemp, R.N. et al.:"Articulated Vehicle Roll 

Stability:Methods of Assessment and Effects of 

Vehicle Characteristics"; TRRL Laboratory Report 

788; Crowthorne, Berkshire 1978 

't45


Specialized Procedures for Preparing the Accident-Avoidance Potential of Heavy 

Trucks 

Psul S. Fancher, 

Arvind Mathew, 

The University of Michigan Transportation 

Research Institute, 

United States 

Abstract 

The ultimate goal of the work described here is to 

improve the safety quality of the transportation of 

goods on highways. Even though the transportation 

missions of heavy freight vehicles have led to a variety 

of axle configurations, suspensions, dolly types, and 

articulation joints, this paper provides a condensed 

summary of fundamental mechanical properties that 

can be used to make first-order estimates of the 

braking and steering performances of these vehicles. 

Specialized analysis procedures, based on basic 

properties of tires, suspensions, brakes, and steering 

systems, are discussed. Thcsc procedures provide simple 

analytical methods for predicting vehicle performance 

in maneuvers associated with operating truck 

combinations on highways. The performance characteristics 

considered are: (l) braking efficiencics, (2) 

transient low-speed offtracking, (3) high-speed otftracking 

in a steady turn, (4) directional and roll 

stability in steady turns, and (5) rcarward amplification 

and roll stability in an avoidance maneuver. 

The emphasis of this paper is on describing the 

features of simplified analytical procedures for estimating 

the accident-avoidance potential available to 

drivers for handling towing units and having trailers 

and semitrailers follow (track) without rolling over or 

exceeding pavement boundaries. 


Since 1970 the Motor Vehicle Manufacturers Association, 

the National Highway Traffic Safety Administration, 

and the Federal Highway Administration in 

the United States have each supported extensive research 

programs that have examined the braking 

capabilities, directional control and stability, tracking 

fidelity, and rollover limits of heavy trucks. These 

efforts have indicated that the heavy truck is a special 

class of vehicle requiring its own test procedures, 

analytical methods, and computerized models for 

evaluating braking and steering performance. As a 

result of these research efforts laboratory tacilities are 

now available for mcasuring the mechanical properties 

of heavy trucks including those properties associated 

with mass distribution and the properties of truck 

tires, suspensions, and steering systems. Simulation 

models are now available for predicting the braking 

746 

and steering responses of heavy trucks. These simulation 

models use mechanical ptoperties measured in the 

laboratory to predict how vehicles will perform on the 

highway or in vehicle tests. Te$t procedures have been 

developed for demonstrating the performance capabilities 

of trucks in maneuvering situations in which 

some heavy vehicles are known to have very limited 

capabilities. The results of tests and simulations have 

been compared to verify that the model builders 

understand the observed phenomena. The current 

state of knowledge provides considerable insight with 

regard to maneuvering characteristics such as: 

r low-speed offtracking (cornering in town) 

r high-specd offtracking (turning at highway 

speeds) 

r braking efficiency (constant deceleration 

braking) 

I roll stability (rollover) 

r steering sensitivity (handling) 

. rearward amplification (obstacle evasion) 

The maneuvering characteristics listed above have 

been selected for use in estimating how well vehicles 

will perform relative to the following practical goals 

that are, in essence, accident avoidance goalst 

' 

r the rear of the vehicle should follow (track) 

the front with adequate fidelity 

r the vehicle should attain a desirable level of 

deceleration during braking 

the vehicle should remain upright (not roll 

over) 

. the vehicle should be controllable and stable 

enough to follow a desired path. 

With respect to these goals, the vehicle should be 

able to perform acceptably over appropriate ranges of 

operational factors including loading, speed, roadway 

friction, lateral acceleration, tire wear, brake maintenance. 

etc. 

Based on the current status of knowledge concerning 

(a) the maneuvering performance$ of heavy 

trucks, (b) analytical proccdures for predicting vehicle 

responses, and (c) experimental procedures for examining 

truck properties, researchers[1,2] have developed 

plans I'or guiding decisions on safety benefits. As 

illustrated in Figure l, the item entitled "definitions 

of response phenomena" is the central factor in a 

research program that would provide information 

needed to weigh the tradeoffs between costs and 

benefits in determining the boundaries between acceptable 

and unacceptable performance. The definitions 

of pertinent respon$e phenomena provide the

foundation for (l) assessing the distributions of performance 

levels existing in thc currcnt truck fleet' (2) 

estimating the links betweeu vehicle characteristics and 

certain types of truck accidents, (3) developing practical 

countermeasures for various types of trucks and 

truck accidents, and (4) developing test methods for 

demonstrating the performance capabilities of specific 

vehicles. These definitions are expected to portray 

response characteristics in selected maneuvering situa' 

tions in terms of performance signatures (that is' 

graphs of performance capabilities), and perforlnance 

measures (that is, distinguishing features of the signatures). 

(Sets of maneuvers, signatures, and measures 

for trucks have been discussed in previous 

publications[3,4].) The purpose of this paper is to 

further the definition of these respollse phenomena by 

describing basic mechanical considerations that have 

been incorporated into specialized procedures (simplified 

analysis-methods) developed for predicting performance 

levels using limited amount$ of paratnetric 

data pcrtaining to the mechanical properties of heavy 

trucks[5]. 

Pertinent Mechanical Properties for 

Making First Order Estimates of 

Accident Avoidance Potential 

The mechanical properties of the components of 

heavy trucks have been corupiled into a 

"factbook"[5]. The influences of component properties 

on maneuvering performance are discussed, and 

Figure 

N}IOWLEDoE EASE FOB GUII'IHG 

BAFETY.RELATEO DECISIONS 

S{lotY 

tlelsled 

1. Knowledge 

quirement$ 

JustitiFd Levsls 

ol PerlormancE 

RsquirBmqdls 

I 

I 

? 

bese for evaluating safety 


tables indicating the relative importance of various 

mechanical properties to the dynamic performatrce of 

heavy trucks are presented there. The following discussions 

relate pertinent mechanical propertie$ to the 

maneuvering situations listed in the lntroduction. 

Low-Speed Offtracking 

The amount of low-speed offtracking for a particu- 

Iar vehicle depends upon its basic dimensional properties 

indicating where the axles are located (how far 

apart they are) and where articulation joints (hitches) 

are located. For simplified analyses, tandem and 

multiple-axle suspensions can be treated as single axles 

locatccl at the centers of the suspension groups. These 

simplifications will not be accurate for vehicles with 

wide spread axles operating on slippery road surfaces, 

but otherwise they are sufficiently accurate for first 

order estimates of offtracking. 

High-Speed Offtracking 

The amount of high-speed offtracking depends 

upon those mechanical properties that influence lowspeed 

offtracking plus pertinent mechanical properties 

describing the installed tires and the distribution of 

loacl. The needed information on load distribution can 

be supplied by specifying axle loads' For simplified 

analyses, the total load otr tandem axle sets will 

suffice if the tandem's suspensions have load leveling 

mechanisms. 

The information on axle loads is used in determining 

the angles that a vehicle's tires must assume to 

provide adequate side forces for the vehicle to follow 

highway curves at highway speeds. An additional 

piece of data needed for estimating these angles is the 

"cornering 

stiffnesses" of the tjres. The angles involved 

here are "slip angles", and measured data on 

thc lateral force capabilities ol' tires ate usually 

presented as functions of slip angle and load. The 

cornering stiffncss is the rate of change of lateral 

force with respect to slip angle evaluated at small slip 

angles and at loads specific to the vehicle's (the tirc's) 

operating conctitiotr. 

Braking Efficiency (Constant Deceleration 

Braking) 

In addition to the axle and hitch locations and 

"static" loads as nceded for the previous maneuvering 

situations, the heights of the centers oi' gravity of the 

units comprising a vehicle are required for analyzing 

braking performance. Also, the heights of the hitches 

are needed to determine the loads existing on the 

vehicle's axles when the vehicle is braking at a given 

level of deceleration. 

To determinc the friction level required to prevent 

wheels frorn locking up and vehicles from becotning 

directionally uncontrollable, the distribution of braking 

effort from irxle to axle is important. Ideally' the 

747

instantaneous ratio of braking force divided by vertical 

Ioad would be the same for all wheels-that way, 

no wheel would require more roadway friction than 

another wheel-this corresponds to the maximum 

efficiency for utilizing a prescribed level of roadway 

friction. However, the distribution of braking forces 

lrom wheel to wheel may not Lre capable of achieving 

high efticiency over the in-use rarlge of road surface 

and vehicle loading conditions. Special equipment, 

such as antilock systems or load sensing proportioning 

systems, are needed to prevent wheel lock arrcl to 

achieve good braking efficiencies- The proportioning 

arrangemcnts provided by conventional braking sys_ 

tems have an important influence on braking efficiency, 

and in order to calculate proportioning, brake 

efl'ectivencss (gains as funcrtions of pressure) are 

needed for each brake. 

Some of the load leveling mechani,sm$ used in heavy 

trucks react to brake torques in a manner that cau$es 

interaxle load transfer between axles in tandem sets. 

This Ioad transfer can be large enough to have a 

significant influence on braking cfficiency. 

Although brake effectiveness or gain seems like a 

straightforward matter, it is not. Brake effectiveness is 

strorrgly influenced by random variations in lining 

flriction, brake wear, work history (temperature history), 

and adjustment and maintenance practices. 

Nevertheless, simplified analyses can provide first 

order estimates of braking efficiencies that are suitable 

for judging the relative performances of vehicles 

that are expected to have comparable levels of mainte_ 

nance and braking experience. 

Rollover 

To make first-order estimates of a truck's rollover 

thrcshold (that is, the level of lateral acceleration at 

which a truck will rollover), it is necessary to treat the 

vehicle as an assembly of sprung and unsprung 

masses. Even so, the most important cleterminant of 

roll stability is magnitLrdc of the ratio tormed by 

dividing half of the track width of the vehicle by the 

height of the center of gravity of the sprung mass of 

the vehicle. The amount of lateral translation of the 

sprung mass is much less imfiortant than the ratio of 

track width to c.g. height, bur it is still critical cluring 

rolling. Suspension roll stiffnesses and roll center 

heights are needed to predict this lateral translation, 

and also, to predict the roll restoring moments tend_ 

ing to keep the vehicle uprighr. 

The vertical stiffnesses of the tires serve to react the 

suspension roll moments against the ground. The 

maximum amount of this restoring moment at any 

axle is limited to the product of half of the track 

width times the axle load- The clistribution of roll 

stiffnesses and vertical loads from axle to axle determines 

those axles whose restoring moment,$ will be 

748 


limited as the vehicle approaches its rollover threshold. 

Finally, the hitches have a significant part to play in 

the rolling process. The typical pintle hitch does not 

provide a roll restraint bctween the units that it 

connects, while fifth wheels and turntables do provide 

roll constraints. The analysis of combination vehicles 

employing pintle hitches rcquires individual roll analyses 

for each independently rolling section of the 

vehicle. 

Steering Sensitivity (Handting) 

The term "steering sensitivity" is associated with 

the amount of steering wheel angle (or front wheel 

angle) required to maintain a steady turn ar constant 

vclocity. It is a performance measure associated with 

*'handling." Specifically, it refers to the amount of 

change in steering angle accompanying a unit change 

in lateral acceleration (that i$, thc inverse of the 

lateral acceleration gain with respect to steering 

input$). When the steering sensitivity approaches zero 

for a particular vehicle, fhat vehicle is approaching a 

situation in which a small change in stccring angle 

cau$es a large change in lateral acceleration. In this 

case, the vehicle is approaching clivergent instability, 

and it will be more difficult to steer properly than 

other vehicles having larger steering sensitivities (that 

is, vehicles having less gain per unit change in steer 

angle). 

Steering sensitivity depends primarily upon how the 

tires on the steered "towing" unit (either a truck or a 

tractor) are loaded and upon the cornering stittnesses 

of those tires. 

To the extent that the properties of towed units 

influence the vertical and lateral loads on the tires of 

the towing unit, the towecl units influence steering 

sensitivity. With regard to steacly turning, information 

on the properties of typical full trailers is not necded 

because their clollies are equipped with pintle hitches 

that do not apply significant vertical loacls. lateral 

loads, or roll moments to the unit ahead of the clolly. 

In addition to the static loading of the towing unit'r; 

tires, the amount of sicle to sidc loacl transfer at the 

various axles of the towing unit is important. This is 

because a large amount of side to sicle loacl transfer 

can cause a small but important reduction in the total 

cornering stiffness of the tires installed on an axle. 

The factors mentioned above imply that in order to 

evaluate steering (handling) performance it is neces_ 

sary to know all of the properties mentioned previ_ 

ously for the other maneuvering situations plus information 

concerning details of the influences of vertical 

Ioads on tire cornering stiffhesses. 

Rearward Amplification 

In combination vehicles, primarily those employing 

full trailers, the motion of the last unit can be a

greatly amplified version of the motion of the leading 

tractor or straight truck. This amplified motion is 

analogous to "cracking the whip", and it is referred 

to as "rearward amplification." Rearward amplificalion 

occurs in rapid (emergency) obstacle avoidance 

maneuvers performed at highway speeds[7]. 

Examinations of (a) results from experimental 

studies[8,9] and (b) the equations of tltltion for 

articulated vehicles[0,11] indicate that the properties 

of both towing units aud towed units have significant 

influences on rearward amplification. 

The important mechanical properties of trailers as 

towed units are their wheelbases and their ratios of (a) 

the sum of all of the cornering stiffnesses of the tires 

installed on the trailer divided by (b) the mass of the 

trailer. 

The important properties of towing units (be they 

trailers, semitrailers, tractors, or trucks) are (a) the 

distance from their center of gravity to the rear hitch, 

(b) the locations ot their wheels, and (c) thc cornering 

stiffnesses of thcir tires. 

Forward velocity has a large influence on the 

amount of rearward amplification. Rearward anrplification 

becomes irnportant at speeds oI approximately 

45 mph (72 kph) for vehicles that are susceptible to it, 

and it becomes increasingly larger as speed incrcases. 

Simplified Analytical Methods for 

Predicting Truck Performance 

The following analysis procedures might be viewed 

as nurnerical equivalents of rules of thumb. They have 

beerr programmed for use on personal computers[5]. 

T'hese calculation methods are based on specialized 

models pertaining to specific maneuvering situations. 

Detailed descriptions of various versions of these 

models will be published in forthcoming reports[5,12]. 

The following discussion is aimed at providing an 

understanding of the basic ideas u.ted in these sirnplified 

analyses. 

Low-Speed Offtracking 

In this case the vehicle may be envisioned as an 

assembly of axles and hitches with specified locations. 

The basic principle involved here is that, at low speed 

near zero velocity, thc wheel planes of thc tire-sets 

will remain tangent to the paths of motion of the 

tire-sets. (That is, the slip angles will be zero.) To 

illustrate this idea consider a simplified unit of a 

combination vehicle in which a generic tire-set is 

respondirrg to thc rnotion of its leading hitch point 

{see Figures 2 and 3). A discrete approximation to tlle 

path of the tire-set can be obtained by the constructiorr 

illustrated in Figure 3. Thc tirc always remains a 

fixed distance behind the hitch point and it proceeds 

along a straight linc segment that is determined by the 

discrete points describing the hitch motion and the 


last position of the tire. (In som€what mathematical 

terms, the hitch point follows a general curve and the 

path of the tire is the "tractrix" of that general 

curve.) 

Given the path of the leading hitch point and 

having determined the path of the tire, one can 

deterrnine the path of a trailing hitch point at the rear 

of the unit. 

In practise, the general curve of the center of the 

front axle of the vehicle is given. (That is, it is 

assumed that the driver steers to attain a desired path 

for the front axle.) In addition, the discrete pctints 

describing the path of the front axle are at small 

intervals-say I foot (0.3 m) apart. Furthermore, the 

path of each hitch point is first calculated using a 

numerical equivalcnt of the graphical constructiorr 

illustrated in Figure 3, and then it is used as the 

general curve for determining the motion of the 

preceding unit and its rear hitch point. 

High-Speed Offtracking 

During low-speed offtracking, if the front axle 

follows a constant radius turn long enough (usually 

more than through 90 degrees of turn), the vehicle 

will eventnally reacl-r a steacly condition with a tixed 

level of offtracking. This level of offtracking can be 

determined from the properties ot' right triangles since 

the tires operate at nearly zero Iateral force, that is, at 

nearly zero slip angles. 

When a vehicle rregotiates a steacly turn at highway 

speeds, its tires generate the lateral forces needed to 

make the turn. The tires must operate at non-zero slip 

angles (see Figure 4) to prodr.rce these lateral forces. 

The geometry of the steady trrrning situation at 

highway speeds differs from the low speed situation 

due to the non'zero slip anglcs developed at highway 

speeds. Figure 5 shows the influence of slip angle, cu, 

on the geometry of a generic trailer. The equation 

given in the figure provides an approximation to the 

Stalionary Oblect Struck by 

Offtracking Trailer 

Path ot lhe 

Olttrackinq 

Figure 2. In low-speed otttracking, each axle tracks 

inboard of the preceding axle 

749

+L,Whelbsc+ DLhE {FlFoxitudon 

to thc BencrEI cuwc 

FpEsfltin8 thE motiod 

ofthc hitch point 

Flgure 3. Offtracking behavior of a $emitreiler following 

the motion of the fifth wheet kingpin 

side force required from the tire set. Using this 

approximation, slip angle is found to be a function of 

cornering stiffness of the tires, vertical load on the 

axle, and the lateral acceleration of the turn. Once the 

slip angle is determined, the analysis is reduced to 

trigonometry; specifically, rwo applications of the law 

of cosines provides the radii of the wheel set and the 

rear hitch point if the center of the front axle is 

assumcd to follow a curve of a given radius. 

As illustrated in Figure 5, there is a speed at which 

the offtracking is zero. Above this speed the wheel 

tracks to the outside of the turn, and below this speed 

the wheel tracks to the inside of the turn reaching 

maximum inboard offtracking at zero speed. Typical 

heavy vehicles operating at highway speeds on highway 

curves tend to offtrack towards the outsicle of the 

turn by a small amount-on the order oF I or 2 feet 

(0.3 to 0.6 m)-but enough ro cause tripping on curbs 

if the driver is not aware of the position of the rear 

axles. 


Roar Arlo Otftrrcldm 

OulMurd ol Frod Axlt€: 

ol Frcnl Axle 

Path ol R€ar A116 

Figure 4. When cornerlng at speed, trailers may offtrack 

outboard ol the tractor if the tire slip 

angfes are farge enough 

750 

,: 

Braking Efficiency 

The braking capability of a vehicle can be estimated 

by analyzing its pcrformance in a series of constant 

deceleration maneuvers. By assuming straight line 

motion, the complexities of vehicle roll and yaw can 

be neglected. 

It the brake forces at the wheels are known. the 

vehicle's mass distribution helps determine its result_ 

ing deceleration. Thc pitching motion of the vehicle 

characterized by loads being transferred from the rear 

wheels onto the front wheels, can be represented in a 

straightfbrward sequential calculation starting from 

the rear of the vehicle. The pitch moment and force 

balances (see Figure 6) for each unit yield hitch and 

axle loads, which when carried forward to the power 

unit, would define the wheel load distribution for the 

vehicle. The amount of load that is transferred 

depends upon the deceleration level, the hitch and 

axle locations, c.g. heights, static load distribution 

and brake proportioning. 

Special consideration is given to tandem sets, which 

redistribute thcir axle loads based upon a percentage 

of the braking effort generated at their wheels (see 

Figure 7). This interaxle load transfer is incorporated 

into the force and moment balance for each unit. The 

Figure 5. lllustration of 

tracking 

It Rw is the rddius subr8nded by the 

a}u,t2. zqJvu mqso -c t, f 

forffill d srd 

zero olflracking, R=fu 

ano nsncE, 

it /57,3,wt/?R 

Va (wbCog57.3/2F.)os 

llole that Vodo€s not 

deFend umn FI. 

c*u, F.v?Hg 

vo, the speed for off-

t+- 

| '.J-" I 

1*ro s 


+J 

,-v,tLr- '4 

tMR? 

rzR2 

Figure 6. Oimensions and freebody diagram lor a Eactor 

semitrailer 

analysis also accounts for full trailers with fixed 

dollies. lnstead of transferring some load onto the 

forward unit, the turlltable of the fixed dolly causes 

much of thc pitching motion to be reacted out at the 

wheels of the full trailer. Besides altering the equa- 

tions for the force and moment balances, the full 

trailer doe$ not complicate the analysis. 

es 

The wheel loads determine the amount of road 

friction required to achieve the cotnputed level of 

deceleration. The friction is given by, F = Fx/Fz. 

The axle whose wheels require the maximum 

amount of friction is most prone to wheel lock-rrp and 

is therefore the critical axle. The braking efficiency is 

defined as the ratio of the vehicle's deceleration to the 

friction required by the critical axle. 

At higher levels of deceleration, rear-to-front load 

transfer increases. Wheel unloading caused by the 

load transf'er increases the frictional demands of the 

vehicle r€sulting in a reduction in the braking efficiency. 

Braking efficiency can therelbre be portrayed 

as a function of deceleration. In addition to deceleration, 

braking efficiency is very sensitive to the vehicle's 

static loading condition. Empty vehicles tend to 

have low efficiencies implying the dangers of locking 

wheels and losing directional corttrol. 

Roll Stability 

The conditions for roll equilibrium are determined 

in this analysis. The vehicle is envisioned as proceeding 

through a sequence of roll angles of its sprung 

NotE!1 

F..'TL + t(FB?+FBr) 

Fzr. + 

thrrc F1rryr Flat F3t 

qnd lhr sytltrfl 

- F*(FE.+F6r) 

l) For o wollriilq bcsm fhr 

lood on ihc hont is inorored, 

ic. F .>O. 

2) Foro4lprinq, P .*fl. 

' 

l+l.l 

V -Ft5 

1 t 

l - l - 

'tza tFzI 

ir rtploctd Fy fhE rytfcm 

I rzil 

- l 

I!?- I 'a:lJMil 

'r F:z( Fsz+ Frr I 

Thc irrm rP12(FsatFgal needi to bt qdded to lhe qppropriqtr 

gifch iror|l€nt tquolion, 

Figurc 7. Approximets representation of interaxle load tranefer

Suspension 

Flgure 8. A heavy truck in a left turn 

Otrt board shift ol CG fforn track cenrer 

mass. For each of these roll angles, the equilibrium 

equations of motion are solved for lateral accelera_ 

tion, the roll angles of each axle, and the relative roll 

angles between the sprung mass and each of the axles. 

(See Figure 8.) These angles dctcrmine the restoring 

moments produced by the suspensions and the tires. 

The conditions for wheel lift-offs are determined and 

the equations are adjusted to account for the ,,satu, 

rfltisn" of restoring moment accompanying lift-off'. 

As the suspensions and tires deflect the mass cenrers 

translate to the outside of the turn, thereby increasing 

the moment tending to produce rollover. The results 

of this analysis are examined for the maximum level 

of lateral acceleration that can be sustained. This is 

the rollover threshold. 

Steering Sensitivity (Handling) 

This analysis is based on the equilibrium equations 

for steady turns at constant velocities. The required 

front wheel $teering anglc is calculated as a function 

of lateral acceleration. In addition, if the vehicle can 

become unstable at high speeds and lateral accelerations, 

the boundary between stable and unstable 

operation is determined. This boundary is displayed 

on a graph with lateral acceleration as the horizontal 

axis and forward velocity as ths verticat axis, thereby 

indicating those combinations of lateral acceleration 

152 


and forward velocity for which the vehicle will be 

stable. 

Examinations of the equations of motion for articulated 

vehicles show that during a steady turn there is a 

single value of yaw rate that applies to all units in the 

combination vehicle-otherwisc, the articulation angles 

would be changing and the motion woulcl not be 

a steady turn. This observation leads to a procedure 

in which lateral acceleration is specifiecl at each step 

of a series of computations made at a constant 

velocity. (In a steady turn, lateral acceleration is the 

product of velocity multiplied by yaw rate.) 

The general form of a step of this procedure is 

illustrated in Figure 9. Applicable notation is defined 

by the figure. The items shown in this figure can be 

used to explain thc basic ideas involvecl in the 

calculation. First, the force of constraint acting at the 

front of the last unit can be calculated from properties 

of the last unit and the selected values of yaw 

rate, r, and forward velocity, u. Next, since the force 

of constraint at the front of one unit has an equal but 

opposite reaction on the next unit, the force at the 

rear of the next unit is known. As indicated in the 

diagram, the force of constraint at the front of any 

unit can be derermined if (a) the force at its rear hitch 

point, (b) yaw rare, and (c) I'orward velocity are 

known. The analysis proceeds in this manner going 

from unit to unit towards the front of the vehicle. 

Once the analysis reaches the forward unit (the 

tractor) the arlount of steer angle needed to obtain 

the required force at the front of the first unit is 

determined, thereby providing the information desired 

for predicting thc stccring sensitivity aspccts of handling 

performance. 

The forces of constraint corrtain information that is 

useful in understanding the handling performance of 

Yrr - Ynr 

Yrs = \hz 

\fr -\fia 

\h4 =0(lastunir) 

Fi iunilE*J.dlpill€|. 

YHi : larceg al mnstainl 

Yt : lorcAs ol mndraint 

Ii : aditulatiof, arut€ 

Xfl ,cut"rc" uv""n ".s. adtw{dHb} 

XCt , Ctstancs U"v""n ".g. atu resr hnd 

Ccl l i trtu#g Erifin€s ot rtu fras an ha tnt sE 

Flgure 9. 

randu Ii-F,-Fz+(/qr-Xr:)i 

Ii- Fe-Fa + (Xq.-X5.)rl 

Ii - Fs- F. * (lhs- xrr)i 

Cot I; + l.Q ; Sln fi - fl 

The elements involved in calculating steer_ 

ing angle, 6

certain vehicles. For vehicles with full trailers the 

constraint force at the front of the dolly is usually 

very small so that changes in the rearward units have 

little influence on the steering response of the tractor 

(or truck). For tractor semitrailers, the lateral force of 

constraint at the fifth wheel is approximately equal to 

the lateral acceleration in g's times the static load on 

the fifth wheel. This piece of knowledge can be used 

to simplify the analysis of the tractor semitrailer, or at 

least, to chesk results from more complicated calculations. 

(See Appendix A for further discussion of 

Figure 9.) 

Rearward Amplification 

In a full trailer, the conventional dolly provides a 

wagon tongue type of steering that causes the center 

of gravity of the trailcr to follow the path of rhe hirch 

point with excellent fidelity in normal lane changing 

maneuvers. However, for sudden obstacle avoidance 

maneuvers with steering inputs having periods on the 

order o[ 2 to 3 seconds, the motion of the trailer'$ 

c.g. can be an exaggerated version of the motion of 

the hitch point. As explainecl earlier, the severity of 

this tendency depends upon the rnechanical properties 

of the trailer and its towing unit. 

In a surprise avoidance maneuver, the full trailer is 

Iike a mechanical serv'ornechanism in which the motion 

of the pintle hitch is the input and the motion of 

the trailer'.s c.g. is the output. The magnitucle of the 

input in the avoidance maneuver depends upon the 

properties of the unit towing the full trailer. The 

magnitude of this input depends also upon the period 

of the maneuver. To fincl the maneuver producing the 

largest amplification, it is necessary to compute results 

using a set of periods covering the range in which 

maximum amplification is expected. Figure l0 illustrates 

the type of result that can occur at a high level 

of amplification. 

In order to avoid performing a number of simulations 

to evaluate r€arward arnplification, arrothcr 

approach has been used to obtain an estirnate of the 

tendency tctwards rearward arnplification. This approach 

consists of evaluating the linear range transfer 

function hetween the lateral acceleration of the leading 

unit and the lateral acceleration of the trailing 

unit. Although not as accurate as the simulation 

approach, the approach using the transfer function 

allows one computation to be used to find the 

maximum amplification. 

Furthermore. the overall transfer function can be 

approximated by products of intermediate transfer 

functions relating the motions of c.g.'s to the motions 

of hitch points and then the nrotions of hitch points 

to the rnotions of c.g.'s in proceedirrg frorn the front 

to the rear of the vehicle. This approximate 

"factoring" 

of the transfer function is possible because 


the forces at the pintle hitches are small enough to be 

neglected. This helps to identity which porrions of rhe 

overall tran$fer functions are the important contributors 

to rearward amplification, and thereby aids in 

identifying where improvements can be made by 

changing vehicle design. 

Summary and Concluding Remarks 

Clearly, a vehiclc's tires are the connections to the 

roadway that are used to achieve the driver's transportation 

objectives. The tires provide (a) forces that 

cause the vehicle to move longitLrclirtally, laterally, and 

directionally plus (b) the forces that support the load 

and hold it upright. The qualities required for maneuvering 

heavy trucks and avoiding accidents in braking 

and stecring situations are depetrdent upon the following 

tire-related matters: 

r where the tires are located on the vehicle 

. how the tire$ are loaded 

r how the tires are braked 

. how the tires are oriented with respect to 

their direction of motion 

Specifically, this paper indicates that values for the 

following meclranical properties are needed to make 

first order estimates ol' the braking and steering 

performance qualities of heavy trucks: 

-locations of the axles 

-locations of the hitches 

-cornering stiffnesses of the installed tires 

-static axle loads that the vchiclc exerts on the 

paveI]tent 

-effectiveness of the brakes installed on each axle 

-interaxle load transfer in tandem suspensions 

P6aI $ffind lraler 

lal€rJ s@l€ratbo. ey4 

Figure 10. In a rapid lane change maneuver, rearward 

amplification results in "crack-the-whlp" 

action of the rear trailer, sometimes reeulting 

in rear trailer rollover 

153

-weights and center of gravity heights of the sprung 

and unsprung ma$se$ of the units comprising the 

vehicle 

-suspension roll stiffnesses (or spring stiffness and 

spread plus auxiliary roll stiffness) 

-roll center heights of each suspension 

-vertical stiffnesses of the tires 

-track width between the centers of the tire sets 

in$tall€d on each axle 

-types of constraints provided by the hitches used to 

couple units together 

-influences of vertical load on the cornering stiff- 

nesses of the in$tall€d tires 

-compliance in the steering systemand 

the steering 

ratio 

Each of these mechanical descriptors has a bearing 

on how the tire$ are operated and how heavy trucks 

will perform in braking and steering maneuvers. 

The analysis procedures described in this paper are 

aimed at evaluating the following performance measure$: 

r offtracking at in-town intersections 

r offtracking on highway ramps 

r friction utilization and braking efficiency 

during deceleration 

r the lateral acceleration threshold at which 

trucks rollover 

r the steering sensitivity and stability of tractor 

semitrailers, single unit trucks, and the leading 

units of multi-articulated vehicles 

I the rearward amplification of the lateral 

acceleration from the first to the last unit of 

articulated vehicles. 

This list of maneuvers could be expanded to include 

matters such as: 

r 

. 

r 

r 

response times in sudden turning situations 

braking during turning 

respon$es to external disturbances (wind 

r 

r 

gusts, road bumps, etc.) 

brake fade during mountain descents 

frictional requirements lor negotiating turns 

on slippery road surfaces 

Nevertheless, the set of performance measures discussed 

herein provides a starting point for safety 

improvement programs of the type illustrated in 

Figure I. 

It is intended that the understanding provided by 

describing and discussing simplified models will assist 

vehicle designers, accident investigators, and highway 

designers in making informed judgements concerning 

the accident avoidance potential of various types of 

heavy trucks. 

7s4 


References 

I. Fancher, P.S., et al. "Heavy Truck Stability: 

Synthesis/Program Plan Development." University 

of Michigan Transportation Research lnstitute 

Report No. UMTRI-96-3, Feb. 19g6 

2,. Leasure, W.A. "Issues in Handling and Stability." 

Truck Safety-an Agenda for the Future. 

SAE P-181, Annapolis, Md., June 1986 

3. Fancher, P.S. "An Evaluation of the Obstacle- 

Avoidance Capabilities of Articulated Cornmercial 

Vehicles." Tenth ESV Conference, Oxford, 

England, August 1985 

4. Fancher, P.S. and Mathew, A. .,Using a Vehicle 

Dynamics Handbook as a Tool for Improving 

the Steering and Braking Performances of Heavy 

Truck$." SAE paper no. 870494, Feb. l9B7 

5. Fancher, P.S. and Mathew, A. ,,A 

Vehicle 

Dynarnics Handbook for Single-Unit and Arriculated 

Heavy Trucks." Final Report to NHTSA 

on contract DTNH22-83-C-07IBT, to be published 

6. Fancher, P.S., et al. ..A Factbook of the Mechanical 

Properties of the Components for 

Single-Unit and Articulated Heavy Trucks." Report 

on NHTSA contract DTNH22-83-C-07I83, 

Dec. 1986 

7. Winklcr, C,8., et al. "Parametric Analysis of 

Heavy Duty Truck Dynamic Stability." Final 

Report NHTSA contract DTNH22-80-C-07344, 

March 1983 

8. Ervin, R.D., et al. "Influence of Size and 

Weight Variables on the Stability and Control 

Propcrties of Heavy Trucks." Final Report 

FHWA contract FH-ll-9577, March l9B3 

9. Winkler, C.8., et al. "Improving the Dynamic 

Performance of Multitrailer Vehicles: a Study of 

lnnovative Dollies." Final Report FHWA contract 

DTFH6I-84-C-00026, July 1986 

10. Mallikarjunarao, C. and Fancher p.S. ..Analysis 

of the Directional Response Characteristics of 

Double Tankers." SAE paper no. ?81064, Dec. 

1978 

ll. Fancher, P.S. "Transient Directional Response 

of Full Trailers." SAE paper no. 821259, Nov. 

1982 

12. Final Report on an MVMA study entitled ,,Development 

of Microcomputer Models of Truck 

Braking and Handling." To be published in July 

1987. 

Appendix A 

The quantities symbolized in square brackets (in the 

form [A,] in Figure 9) contain combinations of tire 

cornering stiffnesses and axle locations having relevance 

to handling considerations. (They also contain 

terms dependent upon the mass of the unit and its

forward velocity.) The important terms related to the 

tires and their locations are (1) the "damping in 

sideslip", that is, the sum of the cornering stiffnesses 

of all of the tires located on the unit divided by the 

forward velocity, (2) the "damping in yaw", that is, 

the sum (over all tires) of the cornering stiffness of 

each tire times the square of its longitudinal distance 

from the unit's center of gravity-all of this divided 

by velocity, and (3) the "corrpling between sideslip 

and yawn', that is, the sum (over all tires in front of 

the center of gravity) of the cornering stiffness of 

each tire times its distance from the center of gravity 

minus the same sum for all tires behind the center of 

gravity of th€ unit-again all of this divided by 

velocity. Without enough damping, vehicles may ei^ 

ther sideslip or yaw excessively to attain the forces 

required for equilibrium. For tractors and trucks, the 

coupling between sideslip and yaw is usually small 

(that is, it is the difference of large, nearly equal 

nurnbers) but nevertheless it is important. If it is 

greater than zero for the towing unit, there exists the 

possibility that the vehicle may be directionally unstable 

at highway speeds. 

The discussion above is independent of the roll 

motion of the vehicle. The vertical loads on the tires 

depend upon thc lateral acceleration of the turn and 

the roll motions of the units. The cortrering stiffnesses 

of the tires depend upon their vertical loads, and 

hence upon the roll properties of the vehicle. Through 

this mechanism, the amount of change in cornering 

stiffnesses is large enough to make some heavy trucks 

directionally unstable at lateral acceleration levels that 

are less than the rollover thresholds of these trucks. 

Another factor of importance is the stiffness of the 

steering system. This stiffness is effectively in series 

with the cornering stiffness of the front tires, thereby 

increasing the amount of steering wheel angle needed 


Antilock Braking Equipment for Heavy Duty Vehicles and lts Evolutions Within 

European Regulation 

J.P. Cheynet, 

P. Beaussier, 

Union Technique de L'Automobile, du Motocycle 

et du Cycle, 

France 

Abstract 

Emergency braking of the heavy duty vehicles, 

especially the combinations, can be dangerou$, particularly 

on wet or icy roads having a low coefficient of 

adhesion. Locking of the wheels having important 

for a specified level of front wheet angte. This is 

equivalent to reducing the cornering stiffnesse$ of the 

front tircs. 

Returning to Figure 9, the a's and b's represent the 

influences of the hitch locations and the influences of 

the forces of constraint on the force and moment 

balances pertaining to each unit. The distance from 

the hitch to the center of gravity of a unit indicates 

the leverage that the associated force of constraint has 

on the yaw motion of the unit, Short lever arms to 

the rear hitch points of dollies generally mean that the 

force of constraint at the t'ront hitch point will be 

small. In addition, for semitrailcrs, if the rear hitch 

point is located near the "neutral force point", the 

force of constraint at the rear hitch has little influence 

on the force of constraint at the front hitch. The 

location of the neutral force point with respect to the 

center of gravity is given by the ratio of the "coupling 

between sideslip and yaw" divided by the "damping 

in sideslip." Although it may not be obvious in all 

cases, there are situations in which the steering 

sensitivities of the leading units of multi-articulated 

vehicles are not strongly influenced by the properties 

of the trailing units. 

Finally, note that the sideslip angles (p's) of the 

motions at the centers of gravity of each unit are 

shown in Figure 9. As indicated by the equations 

given in the figure, the sideslip angles and the yaw 

rate can be used to calculate the articulation angles 

(I's). ln addition to solving for the forces of ct:rnstraint 

at the front of each unit, the equations of 

turning equilibrium can be solved simultaneously for 

the sideslip angles of each unit. Hence, this approach 

could be used for calculating high speed offtracking in 

a manner that includes the influences of spread as 

well as tandem axle sets. 

consequences, lo$s of the trajectory eontrol, jackknifing, 

slipping of the trailer, the equipment and car 

manufacturers have devcloped the antilock devices. 

Some requirements exist since I972 into the regulation 

13 (ONU). The series fitting of the vehicles 

effectively started in 1980 and the experience permit' 

ted to complete these requirements in 1983 (EEC 

Directive 85/6471. 

The requirements ask to keep a good braking 

efficiency, a good stability, a low compressed air 

consumption and no wheel locking. 

755

So the manufacturers have adopted different techniques 

to give the priority either to the braking 

efficiency or to the stability, select high, select low, 

independent regulation. 

Furthermore, the compatibility between tractor and 

trailer shall be studied. 

The wheel speed information available to the onboard 

electronics from the antilock $ystem provides 

the basis for the future introduction of other functions 

such as antislip, retarder control, and speed 

limitation. 

Today, these functions are separated but, for the 

future, it can be supposed that they all will be 

integrated either on the braking system or on the fuel 

injection. 


The ability of a vehicle to come to a stop over a 

short distance depends 

on the following three factors: 

l) The time it takes the driver to react to an 

external stimulus. 

2') The potential braking power of the vehicle, 

3) The friction coefficient between the tires and 

the road surface. 

Reaction time varies widely from driver to driver 

and thus cannot be considered to fall within the 

framework of regulation. 

I3raking power can be verified by a number of tests 

under load. Braking power carnot be put to effcctive 

use, however, unless it is carefully distributed to each 

of the axles and the friction between the tires and the 

road surface is sufficient to make use of it. If this is 

not the case, the $tability of the vehicle can be 

severely impaired by wheel blockage on individual 

axles. The results can be a loss of steerability, 

spinning or, for trailer vehicles, jackknifing. 

ln order, to solve these problems, vehicle and 

equipment manul'acturers have developed equipment 

that can detect the onset of wheel blockage and 

eliminate it by easing braking pressure. 

Principles of European Regulation 

These efforts at eliminating wheel blockage required 

that special stipulations be written into existing regulations 

to ensure that reasonably effective operation was 

maintained. One approach to the blockage problem, 

for example, wa$ to considcrably rcclucc braking 

pressure, which would not have been a satisfactory 

solution. 

It was for this reason that in t970. the United 

Nations Economic Commission for Europe established 

a sct of regulations in this area. This text bccame 

Appendix 13 ol'Regulation 13. 

The stipulations that were adopted cover the fbllowing 

points. 

756 


The addition of antilock equipment must not significantly 

diminish braking effectiveness either on high 

adhesion or low adhesion road surfaces. This must be 

verificd for loaded and unloaded vehicles. 

The equipment must be effective enough to prevent 

thc whecls liom blocking even during hard braking or 

when the vehicle passes from a high adhesion to a low 

adhesion road surface. 

Antilock equipment that is used with brakingenergy 

reserve systems must not itself consume significant 

amounts of power. Braking effectiveness must 

remain sufficient even aller braking continuously tor 

15 to 20 seconds and applying rhe brake pedal lbur 

times in succession. 

Finally, neither the antilock equipment nor the 

braking systcm itself should be adversely affected by 

electrical or electromagnotic ficld$. 

Evolutions 

T'hese regulations enabled European manufacturers 

to market vehicles equipped with antilock equipment, 

which they did beginning in 1980. Four ro five years 

of expericnce with these systems demonstrated the 

need for modifications and additions to the original 

regulations, and in 1985 a working group was forrned 

to do this. This group published EEC Directive 

85/647, which included a new appendix on testing 

antilock equipment. 

The most irlportant points of this new directive are 

as follows. 

- Classification of antilock systems into three categories 

based on perforrnance 

r Category l: must ensure steerability, stability 

and cffectivc braking on split surfaces 

r Category 2: must ensure steerability and 

stability on split surfaces 

. Category 3: must ensure stability on uniform 

surfaces with low as well as high adhesion 

coefficients. 

- Introduction of a split surface te$t that enables 

antilock devices to be categorized as defined above. 

These additions are such that the European countries 

are now prepared to require antilock equiprrrent 

on the types of vehicles included in the following 

table. 

Table 1. 

Buses, GVI{ } I2 T 

Seni-trailer tractors 

GVl4l}16T 

Carrier cspab16 of having a 

trailer and GVW 

) 16 T 

Trailers end semi-treiletB 

GVI{}IOT 

GVW 

r Gross Vehlcle y{eioht 

Category 

Category 

Cateqory I 

ABS on at least two 

wheels on each side

ABS Circuit Arrangements 

Vehicle and equipment manuf'acturer$ have developed 

different techniques in response to diflerent 

directives in the regulations. Thc solutions applied to 

motor vehicles have generally beerr the following (see 

Figure l). 

- One sensor and one trrake pressure control mechanism 

pcr whcel. This system, called independcnt wheel 

control or individual regulating systcm (1.R.), offers 

the advantage of making the maximum use of the 

adhesion potential of uniform road surfaces. It enables 

the maximum braking force to be applied to 

each wheel. The disadvantage of this system is that it 

creates torque that tends to spin the vehicle on its 

vertical axis. When applied to steered wheels, it can 

generate parasitic steering-wheel torque that interferes 

with steering. 

- One sensor for each wheel and one pressure 

controller for both wheels on tlle same axle. The 

pressure threshold for both wheels may set to the 

higher of the two friction coefficients read from the 

sensors: that is called select high regulating system 

(S.H.), but is more usually set to the lower of the two 

friction coefflcients read from the sensors. This is 

called select low regulating system (S.L.). 

The advantage of the select low regulating system is 

that brake drag is equalized, which ensures effective 

steering. The disadvantage is that it does not make 

use of the highest available friction coefficient on split 

road surfaces, with the result that it is less effective 

than the first system (1.R.). 

- f)ne sensor and one pressure controller per wheel, 

but with the additional ability to diminish the differences 

between braking forcesn and thus drag, applied 

to each wheel when the vehicle is braking on a surface 

that offer$ various friction cocfficients. This solution 

repre$ents a compromise between the mo.st effective 

braking and the best stability and this is called 

modified individual system (M.l.R.). 

Heavy vehicle manufacturers have used these solutions 

in concert, applying different solutions to different 

axles. Some examples of what has been done 

(Figures 2 to 4). 

The principles used to outfit motor vehicles with 

antilock have also been applied to trailers and semitrailers. 

The customary configurations are a$ irt figures 

5 to 7. 

Whichever solution is adopted for a motor vehicle, 

it always represents a comprornise between braking 

effectiveness on one hand and steerability and stability 

on the other. 

Further, when connecting a trailer or semi-trailer to 

a motor vehicle, differences in equipment performance 

can give ri$e to cornpatibility problems. A 

notable reflection of this can be seen in the effort.r put 

into improving vehicle coupling. The compatibility 

problem, which has been clearly recognized by manufacturers, 

requires that special care be taken in 

choosing equipment for each of the two coupled 

elements. 

Beginning with the principle of antilock and the 

parameters used in its operation, equiprnent manufacturers 

have also bccn thinking of other vehicle functions 

that could be introduced. 

The primary piece of information is wheel speed. 

What vehicle functions other than antilock de vice 

make use of this parameter? 

' 

t - lndication of vehicle speed by the chronotachograph 

2 - Speed limitation governor 

3 - Drive slip control 

The central nervous system of antilock and the 

functions mentioned above is the onboard processor. 

It follows that rhere is a possibility of making 

important progress in the areas mcntioned, as the 

speed information needed I'or all these functions now 

usually comes from independent sources. 

ELECTRONIC 

BRAKE PILOT SYSTEM ELEMENTS 

Chronotachographs now operate through a mechanical 

or electrical connection to the gear box. The 

speed limitation governor, which acts on the fuel 

injection pump through a control linkage, gets its 

information on vehicle speed from the chronotachograph. 

This information on speed, now used for the speed 

limitation function, could also be used for the antilock 

function. 

This same information on wheel speed could also be 

used in anti slip systems to control the power being 

delivered to the drive wheels when skidding occurs. 

Control would be through a device similar to that 

now used wirh the fuel injection pump to limit speed. 

(See figures I to l0). 

Thc speed information could be used to prompt a 

bricl'braking action on any found to be racing. 

Finally, the retarder considered to be an indispensible 

extension to the braking system should not 

impede antilock operation. French manufacturers 

757

uTAc ANTI-Loct( BBAKTNc EoT..FMENT ffiflm 

Flgure 1 

Flgure 2 

2 AXLES TRUCK 

UTAC ANTI-LOCK BHAKING EOI.FMENT 

3 AXLES TRUCK 


UTAC ANTI-LOCK BRAKING EQUIPMENT 

Flgure 5 

Figure 6 

i-;--------.----r 

- r r : - l 

. q:-----J lF----J L-.:i I 

I 

5 AXLES TRAILER 

UIAC ANTI-LOCK BRAKING EQT.TPMENT 

: q-J 

--l 

l--r I 

I 

=l :l 

L 

3 AXLES SEMI-TRA!LER 

UTAC ANTFLOCK BRAKING EQIJPMENT 

| :=- a-, 

- ' a 

t--:al I 

, €E- -J €l 

' 

I =r f-{ --r 

t _I=t f =t =f ' 

t - B d t i l [ l 

: Y--L :+ ; I 

| [_il____r] 

t 

I 

il _ll_ I 

r l{ frl ?1, 

r i- r i_ir-1 f-___tr 

I L- ts-.- 

: 

_ - J 

5 AXLES SEMI-TRAILER 

OFIVE SLP COfiITHOL 

HNGINE CONTROL

UTAC ORVE SLF CONTHOL 

Figure 9 

DIFFERENTIAL-BRAKE CONTROL 

have done important research in this area, and the 

TELMA Corporation has developed an interface that 

links the retardcr and antilock operations. This interface 

functions as described below. 

When the antilock control processor is actively 

regulating braking through the antilock function, 

retarder operation is temporarily inhihited. 

Once the processor cea$es regulating, retarder braking 

power is gradually brought back to full capacity 

unless the proce$sor intervenes with another comtnand 

to regulate the braking functions. 

This linkage makes it possible to prevent the 

retarder fronr degrading stability on slippery surfaces 

as well as to couple control of the retarder and brake 

operations. 

Figure 10 

COiEilATK I OF DHIVE STf GO'TITROI. 

Alf SFEED COVEHT{OH 


Flgure 11 

Conclusion 

In the near term, it is also possible to imagine 

individual wheel speed sensors linked to a single 

electronic controller that would enable the following 

as they are needed; 

I Speed information feedback for the driver 

r Antilock braking and retarder equipment 

. Speed limitation governor 

r Drive slip control. 

See figure ll. 

The grouping of all these functions will allow a 

diagnostic on the functioning state and a detection of 

pannes. 

See figure 12. 

Figure 12 

ELECTRONIC FUNCTIONS 

I 

7s9


NHTSA'S Heavy Vehicle Brake ResearchProgram-An 

Overview 

Richard W. Radlinski. 

National Highway Traffic Safety 

Administration, 

United States 

Abstract 

The National Highway Traffic Safety Administration 

(NHTSA) has been conducting experiments to 

assess the braking performance ol heavy duty vehicles 

for a number of years. Many different types of 

vehicles have been tested and various aspects of 

braking performance have been evaluated. Experimcnts 

have included full scale vehicle tests on the test 

track, Iaboratory measurements on systems and components 

and dynamometer tests of brake assemblies 

and linings. Tests have included those which measure 

the status quo, as well as those which evaluate 

modifications and hardware designed to improve 

braking performance. 

This paper provides a 

Vehiclc Brake Program, 

areas and presents some 

and conclusions. 

Program History 

The National F{ighway Traffic Safety Administration 

(NHTSA) has been involved in research, testing 

and evaluation of heavy vehicle braking performance 

since the late 60's. NHTSA's first major effort was a 

large scale truck braking study conducted by the 

University of Michigan's Highway Safety Research 

Institute (HSRI). The objectives of this study, initiated 

in 1969 and completed in 1971, were to determine 

the braking perfonnance levels exhibited by 

current design buses, trucks and tractor trailers and to 

establish the maximum braking performance capabilities 

of these vehicles when equipped with difiercnt 

types of advanced brake system hardware. 

Much of the information developed in this study 

was utilized by the Agency in the formulation of 

Federal Motor Vehicle Safety Standard (F'MVSS) No. 

l2l, Air Brake S/s'ferns, which became cffective for 

trailers on January l, 1975, and trucks and buses on 

March I, 1975. 

In 1975 as production vehicles built to comply with 

FMVSS l2l became available, the Agency began to 

evaluate their performance and to compare it to that 

exhibited by vehicles built prior to FMVSS t2l. A 

vehicle test program was established at the NHTSA's 

Safety Research Lab (SRL) in Riverdale, Maryland. 

SRL utilized the track facilities at the U.S. Army 

Proving Ground in Aberdeen, Maryland, flor the 

necessary road tests. Also, as part of this program, 

SRL evaluated stability augmcntation devices (in- 

760 

brief history of the Heavy 

discusses major program 

of the significant findings 

cluding "anti-jackknife" devices) as possible alternatives 

to antilock brake systcms. This initial program at 

SRL was the genesis of the current program at 

NHTSA's Vehicle Research and Test Center (VRTC) 

in East Liberty, Ohio; the SRL is now a division of 

the VRTC. 

SRL's initial program was expanded to address 

various issues being raised as controversy surrounding 

FMVSS l2l began to grow. One of the major 

concerns within the trucking industry was the compatibility 

between the braking systems o[ pre and post 

FMVSS 12l vehicles. Many users reported that mixing 

pre and post FMVSS I21 vehicles in combinations 

(tractor-semitrailers, truck-full trailers, doublcs, etc) 

resulted irr dcgradcd brake performance. ln 1977, SRL 

tested a number of combination vehicles in various 

mixed configurations under a range of operating 

conditions on the test track. Many laboratory tests 

were also run. 

In 1978, two significant events occurred: 1) the 

NHTSA moved SRL from its Riverdale, Maryland, 

laboratory to the Transportation Research Center 

(TRC) in East Liberty, Ohio, to become one of the 

two major divisious within the newly created Vehicle 

Research and Test Center (VRTC) and 2) the Ninth 

Circuit Court of Appeals issued a decision invalidating 

the stopping distance requirements specified in 

FMVSS 121. This eliminated the regulatory need to 

install antilock on heavy vehicles. 

The move of SRL to Ohio slowed the progress of 

the program, but resultcd in SRL having convcnient 

access to extensive facilities ideally suited for heavy 

vehicle testing. 

The court decision had a significant impact on the 

scope and direction of the program. The Agency 

wanted to study the performance of trucks, buses and 

trailers built to conform to FMVSS 121 when antilock 

systems were removed. Also, many different issues 

were raised during the time period that the standard 

was fully in eff'ect that needed to be studicd. With 

improved facilities available and many problcms to 

address, NHTSA established the Heavy Duty Vehicle 

Brake Research Program (as it exists today) in 1979. 

This program, over the years, has addressed many 

different subjects relative to heavy vehicle braking. 

Vehicle road tests as well as laboratory and inertia 

dynamometer tests have been run. In addition to the 

in-house research at VRTC from 1979 to the present, 

the Agency has conducted research on heavy vehicle 

braking systems through contracts with private firms. 

This work includes a three-year in-fleet study of 

automatic brake adjuster performance and reliability, 

and a study of the benefits of retarders for heavy 

vehicles.

The purpose of the discussion which follows is to 

identify the major subject area$ that have been 

addressed in the NHTSA Heavy Vehicle Brake Frogram 

over the years, briefly describe what has been 

done in each area and report some of the more 

significant findings. The references at the end of the 

paper cover the program in more detail. 

Stopping Distrnce and Stability 

During Braking 

Stopping distance tests have been run on over 70 

different heavy vehicles. This group of vehicles con- 

sisted of buses, single unit trucks and combinations 

including tractor-semitrailers, truck-full trailers, doubles 

and triples. Vehicles were tested empty and fully 

Ioaded in straight line stops as well as braking and 

turning maneuvers. Various surfaces from dry pavement 

to ice were utilized. Although some of the tests 

were run with the driver fully applying the brake 

control (i.e., panic application) without limitations on 

wheel lockup and skidding, most testing was done to 

evaluate how quickly vehicles could stop under full 

directional control. This required that the driver 

modulate the brake control to minimize the amount 

of wheel lockup that occurred during the stop. Vehicle 

testing has been performed with fully operational 

brake systems as well as with simulated failures in the 

brake systems. Detailed results of all of these tests are 

given in References l-7. 

Approximately one fourth of the vehicles tested 

utilized hydraulic brake systems; the rest had air 

brakes, the most common system for hcavy vehicles. 

The test results indicate that rhe srable sropping 

capability of the various types of vehicles can be 

ranked as follows: 

Stopping Capability 

Ranking 

1 (best) 

2 

3 

4 

5 (worst) 

Vehicle Type 

Buses (empty and loaded) 

Loaded Tractor Trailers 

Loaded Trucks 

Empty Trucks and Tractor 

Trailers 

Bobtail Truck Tractors 

This ranking is essentially independent of road 

surface coefficient of friction and vehicle speed. The 

ranking applies to '*typical" configuration vehicles in 

these categories in either straight line braking or 

braking while turning maneuvers. 

Looking first at air lrraked vehicles (as currentll 

manufactured), Figure I shows the range of stable 

stopping distances that might be expected from 60 

mph on a straight dry road for the different types of 

air braked vehicles, assuming the brake systems are in 

good condition, burnished and fully adjusted. Performance 

of a typical passenger car is also shown in 

Figure I for reference. 


Buses perform best, primarily becruse under most 

conditions their braking force distribution is close to 

the normal force distribution on their axles. This 

allows buses to achieve maximum utilization of the 

tire/road friction force available at both axles before 

wheel lockup occurs. In effect, the buses have close to 

"ideal" 

braking distribution under most conditions. 

The front to rear weight distribution in a bus generally 

does not change substantially in going from the 

empty condition to the fully loaded condition due to 

the uniform nature of the Ioading. In addition, 

dynamic weight transfer in a bus is low due to a 

relatively low center of gravity height/wheelbase ratio, 

Loaded tractor trailers also perform relatively well 

during braking due to the fact that their braking 

distributions and axle normal force distributions are 

similar. They do not perform quite as well as buses, 

however, due to the fact that the percentage of 

braking on their front (steering) axles is less than 

ideal. Loaded trucks do not perform as well as Ioaded 

tractor trailers. They experience more weight transfer 

onto their front axles than Ioaded tractor trailers 

which cause$ the percentage of braking available at 

the front axles of the loaded trucks to be even further 

below ideal. 

The stable stopping capability of empty trucksn 

tractor,/trailers and, in particular, bobtail tractors is 

relatively poor. This is due to the lact that their 

braking systems, which are sized for the loaded 

condition and have fixed braking force distributions, 

produce too much braking at the rear (or trailer) 

axles. These axles decrease in weight by a much 

greater percentage than the front steering axles when 

the loads are removed. This results in premature 

lockup irnd a corresponding loss of lateral (side) force 

capability at the tires on the "light" axle(s) permitting 

vehicle instability at relatively low deceleration levels. 

ffit* 

'S4tS ffi.,#r*'",* 

-EJ-\ Lodd6d 

|ffi 

Truckg 

#lEmprvrrrcttt 

rrq lTdltrETrolhrg 

r6---EiiJ 

#s 9lS'#' 

Sbpphg olElorlce.{l 

Flguro 1. Stable stopping distance of heavy air braked 

vehlcles from S0 mph on dry siraight road 

76t

The problem is exaggerated if a vehicle has a short 

wheelbase and very lightly loaded rear axle which is 

why bobtail tractors exhibit the worst performance. 

In general, most of the air braked trucks and truck 

tractors tested were found to be "under braked" on 

their front axles in that they would not lock up their 

front wheels before their rear wheels at any load level 

on any of the test surfaces including ice. In addition, 

several of the vehicles were equipped with front axle 

automatic limiting valves (ALV's) which reduce front 

braking substantially when control line pressures are 

low. Since low control line pressures are utilized when 

vehicles are empty, these valves further upset or 

degrade braking distribution in a situation where it is 

already considerably less than ideal. Complete removal 

or deactivation of the front brakes, a practice 

which is common among some truck users, obviously 

degrades the situation even further. The use of ALV's 

or the removal of front brakes increases the chance of 

drive wheel or trailer wheel lockup which can lead to 

spin-out, jackknife, or trailer swing. 

Modification to test vehicles to increase the percentage 

of braking on the front axle such as removal of 

ALV's, increasing the size of brake chambers or 

installing variable brake proportioning systems 

(making braking distribution closer to ideal for 

straight line stops) resulted in optimum performance 

in the braking and turning case. Much shorter and 

more stable stops resulted in both cases. Increasing 

the front brake torque, however, did increase steering 

wheel pull when a vehicle was braked on an uneven 

coefficient of friction surface (difference in slipperi' 

ness left to right). This increase was insignificant if 

the vehicle was equipped with power steering and the 

steering axle had a low kingpin offset (scrub radius) 

but was quite significant when the vehicle had manual 

steering and a high kingpin offset. This indicates that 

steering $ystem design must be taken into account if 

consideration is given to increasing front brake torque 

Ievels substantially above those which now exist. lt 

may be necessary that vehicles have power steering 

and/or better steering geometry if greater levels of 

front brake torque are utilized. 

Current model heavy hydraulically braked trucks 

were found to perform somewhat better than air 

braked trucks.* Figure 2 shows the relative performance 

of typical trucks with air and hydraulic brakes. 

Performance of the hydraulically braked vehicles is 

better primarily because they are typically designed 

with higher torque front brakes and achieve better 

braking force distribution particularly when empty' 

.Class 6 and 7 singlc unit trucks and school bttscs are available from some 

manufacturers with either air or hydraulic brakes. Hydraulic brakes are 

standard on these vehicles and air brakes are oflered as an option. Because 

of thc irdditional complexity of the air brakc system the cost of such a system 

is higher. 

762 


100 e00 300 

STOPFINB D]STANEE. ft 

r.lFAilBE FOF 

L I TYFICAL VEHICLEB 

Flgure 2. Relatlve performance of truckt equlpped 

wlth alr and hydraullc brakee-60 mph, dry 

road 

One point that should be made about the above 

discussion of stopping capability is that it is based on 

the premise that brakes are in good working order, 

burnished and fully adjusted. If this is not the case, 

total brake force output may not be sufficient to 

produce a very high deceleration when the vehicle is 

loaded, even if the brakes are fully applied. With 

degraded output brakes, higher loads result in poorer 

performance, just the opposite of the situation as 

discussed above where the fully loaded vehicles out 

performed the empty vehicles. 

Antilock 

NHTSA tested a number of vehicles (five power 

units, four trailers and one converter dolly) with 

production antilock systems between 1975 and 1977. 

These tests described in References 2 and 3 were 

primarily designed to evaluate the braking performance 

gains provided by the antilock systems and to 

evaluate compatibility of tractors and trailers with and 

without antilock in various "mixed" vehicle combinations. 

Straight line stops as well as braking and 

turning maneuvers were run with both empty and 

loaded vehicles on surfaces ranging from dry asphalt 

to wet Jennite (pavement sealer). Although these tests 

did not specifically include an evaluation of reliability 

and vehicle test mileage was generally low compared 

to typical truck user mileage, the operation of a group 

of antilock equipped test vehicles did provide some 

insight into the problems being reported by the truck 

users. Component failures, electrical connector and 

wiring problems, intermittent failure warning light 

operation, etc were experienced on some test vehicles. 

When the systems were operating properly, however, 

braking performance gains were significant. Vehicles 

stopped shorter and were much more controllable 

with the antilock in operation. Antilock on the front 

axle prevented loss ol' stcering, on the drive axle(s) 

prevented jackknifing and on the trailer axles prevented 

trailer swing. In terms of compatibility between 

vehicles with and without antilock there were

no cases found where havlng antllock on one vehicle 

or one zurle in a combination degraded performance. 

In fact, just the opposite was found. The application 

of antilock to any axle provided a braking performance 

improvement; the more axles that had antilock, 

the greater the improvement. This was found to be 

true with single unit vehicles, tractor semi trailers and 

doubles combinations. 

After the Court struck down the stopping distance 

requirements in FMVSS l2l in 1978, very few trucks, 

trailers, and buses were built with antilock systems 

and very few users attempted to keep existing systems 

operational. Use and production of antilock essentially 

dropped to the negligible level in the U.S. ln the 

meantime, however, interest in antilock began to grow 

out$ide the U.S., particualrly in Europe. Several 

European component manufacturers began producing 

second generation antilock systems and European 

vehicle manufacturers began offering them as optional 

equipment. Thousands of these systems are now in 

use in Europe and although no countries currently 

require antilock, several are considering issuing regulations 

mandating the systems on heavy vehicles. 

Two fleets in the U.S. are currently operating a 

small number of vehicles with one particular brand of 

European antilock. In early 1986, one of these fleets 

made a two-axle straight truck available to NHTSA 

for testing at VRTC. Reference 8 provides a detailed 

description of the NHTSA test program on this 

vehicle. Figure 3 shows a summary of the test results 

and indicates the percent improvement in stable stopping 

distance that resulted when the antilock was 

operational in various types of braking maneuvers. 

Figure 3 is based on comparing the performance of a 

skilled test driver modulating the brakes on the vehicle 

without the antilock to the performance of the antilock 

system. Performance improvement with the antilock 

for more typical drivers would be expected to be 

even greater than that shown in Figure 3. ln addition, 

Figure 3 assumes the driver has the presence of mind 

at#lt Ard.lr foli!h.d 

tuncr.t. 

Flgure 3. Percent lmprovement in controlled/steble 

otopplng distance with antilock-fulfi loaded 

vehicle 

l 

tm,ail/Y 


E** 

I** 

W*^ 

to modulate the brakes if the wheels bcgin to lock. In 

an actual emergency on-road situation, this may not 

be the case: the driver may panic and lock the wheels 

skidding out of control. 

In addition to this two-axle truck, VRTC has 

recently completed tests on a European tractor-trailer 

combination equipped with antilock and is currently 

in the proces$ of testing a U.S. tractor-trailer combination 

with antilock. In the current tests, VRTC is 

evaluating the performance of a 'rstandard" European 

system as well as modified versions of the 

standard system with different (simplified) control 

strategies. The purpose of these tests i$ to quantify the 

performance gains provided by the different control 

strategies. Individual wheel control, axle-by-axle control 

and tandem control strategies are all being 

evaluated. 

In order to evaluate reliability of current antilock 

systems, NHTSA is planning to conduct (under contract) 

a two year fleet study of approximately 200 

vehicles with various available antilock systems. This 

program will start in about a year. Action to find a 

suitable on-board data acquisition system has already 

been initiated. This data acquisition system will be a 

relatively simple device not unlike commonly used 

electronic tachograph and will monitor antilock function 

while the vehicle is in operation in the fleet. The 

NHTSA will also follow (via a contractor) experience 

with antilock in Europe as well as Australia where 

there are also many systems in-use. In addition, the 

NHTSA will attempt to obtain as much information 

as possible on antilock equipped vehicles operating in 

the U.S. that are not specifically included in the 200 

vehicle fleet study. 

Tractor and Trailer Compatibility 

As mentioned earlier, compatibility between vehicles 

with and without antilock in combinations was studied 

by NHTSA in the late 70's. The subject of tractor 

and trailer brake system compatibility, however, is 

much broader than the issue of mixing vehicles with 

and without antilock. In fact, compatibility is a major 

area of concern today even though vehicles do not 

typically utilize antilock. ln general, compatibility 

refers to the braking system ou the tractor and the 

braking system on the trailer working together in 

harmony to provide desirable combination vehicle 

brake system durability antl braking performance. It 

is primarily determined by the transient and steady 

state brake force distribution existing at the various 

axles of thc combination vehicle. 

Both of these aspects of compatibility have been 

addressed in depth in the NHTSA research program 

over the years. The transient brake force distribution 

is strongly influenced by the flow characteristics or 

timing of the pneumatic system. Most of the NHTSA 

763

Trollor Brokgs 

Truclor Slow /Troller FoBl 

(go,00olb Vehicle) 

Tlme 

Figure 4. Klngpln force versus tlme for e panlc stop 


research on pneumatic timing is described in Reference 

9. The time that it takes for the pressure at the 

brakes to reach a particular level after the pedal is 

applied is known as the apply timing; the time that it 

takes the pressure to be reduced to a specified low 

level is known as the release timing. Overall apply 

time is important because it determines how quickly 

full braking force is achieved; this has an influence on 

stopping distance. Relative apply time for tractors 

with respect to trailers is also important because it 

affects the coupling or "kingpin" force between the 

two units during braking. High coupling force is 

undesirable because it aggravates the jackknife situation. 

If the trailer brakes apply slow with respect to 

the tractor brakes, the trailer will tend to ,,bump" Irnolr 

lmrve 

ffi 

the 

tractor harder. Figure 4 shows results of actual 

measurements made on an 80.000 lb combination 

during a panic stop on dry pavement using a semitrailer 

with an instrumented kingpin. It can be seen in 

Figure 4 that when the tractor brakes apply fast with 

respect to the trailer brakes, overshoot in kingpin 

force occurs. This overshoot reaches the same level as 

the case where the trailer brakes are not working (i.e., 

infinite trailer brake apply time). One point that 

should be made relative to Figure 4 is that even with 

the ideal case (trailer applies before the tractor) there 

is a substantial force at the kingpin. 

Recent test$ of typical late model vehicles(Reference 

l0) indicate that trailer brakes do not usually apply 

before tractor brakes. Figure 5 shows apply times (0 

to 60 psi) for nine tractor-semitrailers and six doubles 

combinations. 

Release timing is also important to vehicle stability 

although it has no effect on $topping distance. If a 

driver is applying his brakes in an emergency situation 

and locks the wheels, it is important that he be able to 

release the brakes as quickly as possible; otherwise he 

may skid out of control. Slow release times also affect 

brake temperature and wear. 

Steady state brake force distribution is very important 

to compatibility. NHTSA research in this area is 

covered in References 4 and 10. In sublimit braking 

rnarren 

4 5 8 

VEHICLE No, 

Flgure 5a, Apply time at each axle set*nine tractor 

semitrailere 

situations (i.e., well below the point of wheel lockup) 

brake forces must be balanced, otherwise excessive 

wear and temperature build-up will occur at the 

"overbraked" 

axle. ln limit braking situations, wheels 

on overbraked axles will lock up and skid prematurely. 

The input level at which braking force start$ to 

occur at each brake, known as the brake force 

threshold pres$ure, is a very critical parameter to 

compatibility. If brake force at the tractor starts to 

occur at an input level below that needed for the 

trailer braking to start or visa versa, brake temperature 

imbalance is probable in repeated or continuous 

low pressure braking situations (such as mountain 

grade descents). lf the vehicles are on low coefficient 

of friction surfaces, wheel lockup can occur prematurely. 

Figure 6 shows the final brake temperature on a 

tractor and trailer at the end of a 5 mile long 4go 

grade descent at 45 mph as a function of threshold 

pressure difference between the tractor and trailer. 

Only a 2 psi diffcrence in threshold pressure can make 

a difference of over 200oF in final brake temperature. 

Figure 7 shows the effect of threshold pressure 

difference on braking efficiency by comparing a 

combination with equal thresholds to one when the 

7&4 

sEc0N05 

!rnmr 

0,8 

lonrve 

ffi rnlrrrn r 

[.]oor,r v 

0,5 N 

ffiNffi 

rnaten a 

ffi 

!? L3 

VEIiICLE IIO. 

Flgure 5b. Apply tim6 at each axle set-six doublos 

combinatlons

e0 

E 

d| Eou 

E 

L 

F*o 

H ,oo 

1 1 3 4 5 

tIRSSEolJl PR$93. DIF ERBNCB.FTI 

Figure S. Brako temperatures for 4tf grade elmulatlon-flve 

mlles at 45 mph 

trailer threshold is 4 psi lower than that on the 

tractor. On low mu surfaces particularly with the 

empty vehicle, the drop in efficiency that occurs when 

threshold pressure is different is quite significant,. 

Brake Adjustment 

Both vehicle and inertia dynamometer tests have 

been run to determine the effect of adjustment on 

E 

f 

f 

I 

e 

I 

t 

n 

C 

U 

E rft 

J 

8e. 

6i. 

f 8 c 

I 

2e; 

c 

I 

t 

n 

6 c 

G 

u r . 

e.1 I 

0,c 

e,a 

\ 

i--__ 

TADEH 

r t t l 

e.r e.6 t,8 1.6 

Equat Thrssholds 

Trat lar ,lpot Lou 

UHLADEH 

o,a G.e Q,4 C.6 f .8 t.e 

Frtk TlrazRord Frlctlon (rul 

Flgure 7. Breklng elficiency tor tractor-semitraller 

combination with equal threshold pressures 

and 4 psi I'low" trsller threshold (or 4 psl 

"high" tracter threshold) 


Percent of Fully 

Adjusied Torque 

(ot 2OO"F) 

too 

\ 

Mfgr's Recommerd€d 

Adjustment Ronge 

Figure 8. S-cam drum brake performance ae a function 

of adlustment level and drum tempetatu 

re 

braking system performance. This work is described 

in Reference.s ll and 12. It has been determined that 

the torque output of air braked heavy trucks is very 

sensitive to brake adjustment level. This is not the 

case for hydraulic brakes used on heavy trucks and 

most hydraulic brakes on cars and trucks are of the 

automatic adjusting type. The majority of truck air 

brake systems must be manually adjusted. 

Figure I shows the effect of brake adjustment on 

the output of a typical heavy duty air brake at two 

different temperature levels, 200oF and 600oF (temperature 

in the brake drum). 

The lower temperature repr€sents a relatively 

"cool" 

brake that has not been exposed to a great 

deal of repeated or continuous braking. The higher 

temperf,ture represent$ a relatively "hot" brake, and 

is typical for a mountain descent although it is by no 

means the maximum temperature that a brake might 

experience in service. Figure 8 is for an S-cam drum 

type brake, used on the majority (over 90 percent) of 

heavy duty air braked vehicles. 

Adjustment level in Figure I represents the stroke 

of the air brake actuator (chamber) when the pressure 

in the actuator is 100 psi and the brake is at ambient 

temperature. Normally for the brake shown, the 

stroke of the actuator at 100 psi with the brake fully 

adjusted is approxirnately I.5 inches; this stroke is 

required to take up the slack and deflection in the 

system. As the brake shoe wears, the stroke increases 

\ t\\ 

t \\tI 

e.5 

Adjustment Level(in.) 

ZOO"F Drum 

\ t 

b 6oo"F Drum 

765

due to the greater actuator travel necessary t6 move 

the brake shoes out against the brake drum. For this 

particular brake, the manufacturer recommends that 

the brake be readjusted when the stroke reaches 2.0 

inches although the actuator actually has a full travel 

of approximately 2.5 inches. 

It can be seen from Figure 8 that at 200oF brake 

temperature, brake torque continually drops as adjustment 

level degrades from the fully adjusted level. This 

is true even over the manufacturerts recommended 

adjustment range; at the recommended readjustment 

point (2.0 inches) the torque has dropped to 85 

percent of its fully adjusted level. When the brake is 

hot (600"F), there is a drop to 85 percent even when 

the brake is fully adjusted. This drop is due to two 

factors: l) brake lining fade at the elevated temperature 

and, 2) brake drum expansion which results in an 

actuator stroke increase. Brake torque is reduced to 

50 percent compared to a fully adjusted cool brake, 

when adjustment reaches the mauufacturer's recommended 

readjustment point. This is a signifiicant drop 

even though brake adjustment is considered to be 

acceptable in terms of the manufacturer's recommendations. 

Under this condition, the brake can only 

develop on half of the torque it could ii it was fully 

adjusted and cool. Beyond the manufacturer's recommended 

adjustment range brake torque drop is even 

more dramatic, particularly if the brake is hot. 

Reduced brake torque due to brakes being out-ofadjustment 

affects the brake force balance and overall 

braking capacity of the vehicle. As a result, not ouly 

is limit performance stopping ability affected, but 

downhill operations also become more prone to brake 

fade and runaway. 

Figure 9 shows the results of limit performance 

stopping distance tests conducted on a fully loaded 

6x4 truck at two different adjustment levels: l) fully 

adjusted, and 2) at the manufacturer's recommended 

readjustment point. Both cool brakes (200"F) and hot 

brakes (600'F) are shown. Beyond the manufacturer's 

recommended adjustment range the stopping distance 

tm E06 300 400 !00 @0 ?00 d,0 

Figure L Stopplng distance of fully loaded truck at 

two brake adiustment levels (60 mph-dry 

road) 

766 

EXPERTMENTAL SAFETY VEHICLES 

of the vehicle would be even longer than that shown 

in Figure 9. 

Brake adjustment primarily affects the stopping 

capability of trucks when they are loaded; this is 

where maximum brake torque is needed to decelerate 

vehicle mass. With an empty vehicle, more than 

enough brake torque is usually available to lock the 

rear wheels despite the level of adjustment, unless 

adjustment is so poor that practically no torque is 

generated. 

Retarders 

Research performed under contract by the University 

of Michigan to evaluate the benefits of retarders 

for heavy vehicles(References 13,14,15) indicates that 

these devices can extend brake life and reduce the 

possibility of runaways on downgrades. VRTC, working 

in cooperation with the University of Michigan, 

conducted full scale vehicle tests to determine what 

effect these devices have on vehicle stability and 

stopping performance in limit braking maneuvers. 

Reference 16 describes this effort. The results of tests 

on two different combination vehicles indicate that in 

limit braking maneuvers, retarders can increase the 

stable stopping distances. Since most U.S. vehicles are 

"overbraked" 

on their drive axles, retarders (most 

commonly used retarders act through the drive axle) 

tend to upset brake force distribution even further. 

With retarders in operation, drivers must modulate 

the service brake control to an even greater degree to 

avoid wheel lockup and jackknife. Table I shows test 

results for braking tests on wet Jennite curves. Both 

of the vehicles use "engine brake" type retarders. 

Use of retarders without applying the service brakes 

can also affect vehicle stability. Since retarders, in 

effect, utilize longitudinal friction at the tire/road 

interface, they reduce the lateral friction available for 

cornering. The maximum safe speed for entering a 

curve is reduced whcn the retarder is "on". 

In 

addition, loss of control at the limit speed can change 

from a stable "plow out" mode with the retarder 

"off" 

to an unstable jackknife mode if the retarder is 

'*ontt. 

In order to warn retarder users of the potential for 

stability problems, NHTSA has prepared an infbrma- 

Table t. Effoct of reterdeE on stable stopping dis. 

tancB In sllppery curyes. 

vchtcls ldsdtnr 

6x4-3? Enpty Tratler 

Efrpty Trsllsr 

Bobtstt 

4x2-sl Enpty TrslI€r 

Empty Tratler 

Londcd lrstlcr 

CurvG 

Radtug 

( fr.) 

200 

500 

200 

200 

500 

?00 

Best In-kne 

Inltlal 

sp€ed 

(nrh) 9,/0Rclard+r U/tret{rder 

25 

40 

25 

88 95 

3lr 123 

98 t23 

88 98 

183 224 

98 tO3

tional booklet(Reference 17) and given it widespread 

distribution in the trucking community. This booklet 

encourages the installatiorr of retarders on vehicles 

since they do offer safety benefits and can extend 

brake system tife significantly. The booklet, however, 

cautions against use of retarders (all types are usually 

controlled by an in-cab switch) in situations where the 

vehicle is empty and/or the road slippery. 

Anti-Jackknife Devices 

ln 1975, NHTSA tested $everal different anti- 


jackknife devices on several different combination 

vehicles. Brake-in-a-curve, brake-during-a-lane-change 

and straight line braking tests were run on several 

different surfaces. Although these devices restrict 

articulation and prevent the tractor from hitting the 

trailer, they do not prevent wheel lock, and thus do 

not cure the basic instability problem' Figure l0 

shows a tractor trailer both with and without an 

anti-jackknife device. Without the anti-jackknife device, 

the vehicle could jackkniie if the tractor wheels 

Iock. With such a clevice the vehicle will not jackknife 

but could spin as an entire unit (possibly across 

several Ianes of traffic). In effect, the combination 

becomes a "long" straight truck. 

Anti-jackknife devices will improve vehicle stability 

if the trailer wheels are kept rolling'+ In this case, the 

trailer acts as a "rudder" Figure 10. Loee of stability with and without antl' 

iackknife device 

the trailer cannot be sensed by most tractors' Even if 

for the combination vehicle. the trailer has a large leak, the tractor worlld still be 

Most U.S. vehicles when empty, however, have their able to maintain full reservoir pressure and would 

brakes balanced such that the first axle to lock on thus not give the clriver any indication of the problem 

increasing brake input are those on the trailer' In a on the trailer. Unfortunately, in this catie, the tractor 

panic situation, full application of the brakes almost low pressure warning light and buzzer which senses 

always locks the trailer brakes' It, therefore, cannot tractor reservoir pressure is no help in the event of 

be assumed that the trailer wheels will be rolling in trailer failures. 

limit or emergency brakirrg lnaneuvers. 

Trailer Emergency Systems 

A number of tests have been run to determine if the 

pneumatic systems tln trailers could be simplified. 

This work, describetl in Relerence 18, was performed 

in lieht of comments on a notice of proposed rulemaking 

to modify thc requirements of FMVSS No' 

l2l. Full scale vehicle tests, inertia dynamometer tests 

and laboratory tests of trailer pneumatic system 

"mockups" were performed. The results of these tests 

indicated that simplification would be possible but 

that care must be taken to avoid systems that pernlit 

spring (parking) brake drag withoLtt warning the 

driver. lt was found that drivers could not feel spring 

brake drag even though it was occurring to the point 

of overheating and rcducing the effectiveness of the 

brake system. It was also found in these tests that 

typical tractor plumbing is so restrictive in its delivery 

of supply air to the trailer that pneumatic failures on 

tsote pto-otetg of thac devicet demonstratc them in tests with the trailer 

brakes turned off, an unrealistic operating condition 

Wlthouf 

Anti - Jockknife 

Device 

Performance of U'S, Versus European 

Vehicles 

Comparative testing of a European tractor semitrailer 

built to meet European brake standards (and 

not equipped with antiloc-k) and a U.S. combination 

of equivalent siee and weight has been performed' 

Many different types of braking maneuvers (straight 

line, curves, and lane changes), various surfaces and 

different vehicle loads have been included in these 

tests. Since this work has only recently been completecl 

the report describing this effort in detail has 

not yet been published. The basic conclusion reached 

from these tests, however, is that the braking performance 

of the European combination is generally 

superior to that of the U.S. combination although the 

difference in performance was much smaller than 

expected. The European vehicle had the samc brake 

on the front axle as on each rear axle and load 

sensing proportioning systems on the drive and trailer 

axles ancl this provided a more optimum brake force 

distribution over the range of operating conditions' 

't67

The only tests where performance was greatly different, 

however, wcre those with the bobtail tractors_ 

the European tractor stopped much shorter and was 

easier to control during braking. When a simple 

bobtail proportioning sysrem+ was added to the U.S. 

vehicle, however, the performance of the two tractors 

was essentially the same. 

Brake Linings 

During the last year, dynamometer tests have been 

run to investigate the performance of heavy vehicle 

brake linings. This research which is expected to 

continue for at least another year is addreising two 

issues: I) Iining performance variability, Z) dif't'erences 

between asbestos and non-asbestos linings. Lining 

performance variability is important because it deter_ 

mines how closely brake force balance and braking 

efficiency can be controlled. This impacts tractor and 

trailer compatibility as well. 

Understanding the performance characteristics of 

non-asbestos Iinings is important because they may be 

the only type of linings available in the future. Many 

vehicle manufacturers are now using them and EpA 

has proposed the complete elimination of asbestos in 

brake linings. 

Data from this research effort will be made avail_ 

able to the SAE subcommittee rhar is currently 

developing new test procedures and rating schemes for 

brake linings. The SAE subcommittee is ittempting to 

replace current SAE Recommended practices for 

brake linings that are known to be inadequate. The 

data will also be provided to the SAE subcommittee 

that is developing Recommended practices for tractor 

and trailer brake system compatibility. 

Summary and Conclusions 

NHTSA has heen conducting research on heavy 

vehicle braking since l969. Over the years many 

vehicles have been tested and many issues have been 

addressed. As a result of this effort, the braking 

performance characteristics of heavy vehicles are well 

understood and perfotmance deflciencies have been 

identified. There is a large gap between the perfor_ 

mance of passenger cars and heavy trucks and al_ 

though this gap may never be completely bridged, 

significant improvements are possible. 

References 

l. Murphy, R.W., Limpert, R. and Segal, L., 

"Bus, 

Truck, Tractor-Triler Brking System performance, 

Volume I or Z: Research Findings," 

Final Report, University of Michigan Highway 

Safety Research Institute, DOT NHTSA Contract 

Number FH-l l-7290. March 1971. 

fThis system is available es en option on 6ome U.S. tractors. It reduces 

pressure to the drive axle brakes whcn a trailer i$ not connected td lhc 

tractor, 

768 


Z. Radlinski, R.W., .,Air Braked Vehicle perfor_ 

mance: FMVSS No. l2l Braking Systems Versus 

Pre-FMVSS No. l2l Braking Systems and Sta_ 

bility Augmentation Devices," U.S. Department 

of Transportation Report Number DOT HS_gOl 

967, Augusr 1976. 

3. "Technical Assessment of FMVSS l2l_Air 

Brake Systems," A Report of the FMVSS l2l 

Task Force, U.S. Department of Transportation, 

February 24,1978. 

4. Radlinski, R.W. and Williams, S.F.. ,.NHTSA 

Heavy Duty Vehicle Brake Research program 

Report No. l-Sropping Capability of Air 

Braked Vehicles," Volume l-Technical Report, 

Report No. DOT HS 806 738, April 1g85. 

5. Kirkbride, R.L. and Radlinski, R.W., ,.NHTSA 


Report No. 4-Stopping Capability of Hydrauli_ 

cally Braked Vehicles-Volume l: Technical 

Report," National Highway Trafl'ic Safety Ad_ 

ministration, Report Number DOT HS 906 860. 

October 1985. 

6. Radlinski, R.W. and Flick, M.A., ,,A Demon_ 

stration of the Safety Benefits of Front Brakes 

on Heavy Trucks," Vehicle Research ancl Test 

Center Final Report No. DOT-HS-807 

061, De_ 

cember 1986. 

7. Garrott, W.R., Cuenther, D., Houk, R., Lin, 

J., and Martin, M., ,.Improvement 

of Methods 

for Determining pre-Crash parameters From 

Skid Marks," National Highway Traffic Safety 

Administration, Final Report, Report Number 

DOT HS 806 063, May t98t. 

8. Radlinski, Richard W. and Bell, Steven C., 

"NHTSA Heavy Duty Vehicle Brake Research 

Program Report No. 6-performance Evaluation 

of a Production Antilock System Installed on a 

Two Axle Straight Truck," Vehicle Research 

and 

Test Center, Reporr Number DOT HS g06, 

August 1986. 

9. Radlinski, R.W. and Williams, S.F., *.NHTSA 


Report No. S-Pneumatic Timing," Vehicle Re_ 

search and Test Center, Report Number DOT 

HS 806 897, December 1985. 

10. Radlinski, R.W. and Flick, M.A., ..Tractor and 

Trailer Brake System Compatibility," SAE pa_ 

per Number 861942, November lgg6. 

It. Radlinski, R.W., Williams, S.F. and Machey, 

J.M., "The Importance of Maintaining Air 

Brake Adju$tment," Society of Automotive Engineers, 

Paper Number 821263, November 19g2. 

12. Radlinski, R.W. and Williams, S.F., ,.NHTSA 


Report No. Z-The Effect of Adjustment on Air 

Brake Performance," National Highway Traffic

Safety Administration, Report Number DOT HS 

806 ?40, April 1985. 

13. Fancher, P., O'Day, J., Bunch, H., Sayers, M. 

and Winkler, C., 

14. 

15. 

"Retarders for Heavy Vehicles: 

Evaluation of Perforrnance Characteristics and 

In-Service Costs-Vol. I: Technical Report," University 

of Michigan Transportation Research Institute, 

Report Number DOT HS 805 807, February 

1981. 

Fancher, P.S., O'Day, J. and Winkler, C.8., 

"Retarders 

for Heavy Vehiclesl Phase II Field 

Evaluations," University of Michigan Transportation 

Research Institute, Report Number DOT 

HS 806 297, June 1982. 

Fancher, P.S. and Winkler, C.8., "Retarders 

for Heavy Vehicles: Phase III, Experimentation 

and Analysis; Performance, Brake Savings, and 


Vehicle Stability," University of Michigan 

Transportation Research Institute, Report Number 

DOT HS 806 672, January 1984. 

Fancher, P.S. and Radlinski, R.W,, "Directional 

Control of Retarder Equipped Heavy Trucks 

Operating on Slippery Surfaces," SAE Paper 

No. 831788, November I983. 

"A 

Professional Truck Driver's Guide on the 

Use of Retarders," National Highway Traffic 

Safety Administration, Report Number DOT HS 

806 675, January 1985. 

Radlinski, R.W. and Williams, S.F., "NHTSA 

I6. 

t7. 

18. 

Heavy Duty Vehicle Brake Research Program 

Report No. 3-Evaluation of Parking and Emergency 

Pneumatic Systems on Air Braked Trailers," 

Vehicle Research and Test Center, Report 

Number DOT HS 806 757, May 1985.

ESV.Eleventh.Int.Conf.SECTION.FOUR.Session.Six

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?