Page 2 Lecture Notes in Computer Science 2865 Edited by G. Goos ...

38 L. Qin and T. Kunzthese two protocols show that the signal strength fluctuations cause routes to be assumedbroken, and in turn the packet delivery ratio significantly decreases and packetdelay increases. We apply a new signal power threshold during the route discoveryprocess, so that only the routes with strong signal strength will be chosen. For DSR,we also reduce some control messages to reduce traffic. Our simulation results showthat this can significantly increase the packet deliver ratio and deceased packet latencyfor both protocols.An overview of free space, two-ray ground and shadowing propagation models areprovided in Section 2. Section 3 analyses the impact of a shadowing model on theperformance of routing protocols, and presents our proposal. Section 4 contains thesimulation results under different mobility patterns and parameters. Section 5 drawsour conclusion and directions for future research. A more detailed introduction toDSR and AODV can be found in [3] [5].2 Propagation ModelsThe free space propagation model is used for the situation when the transmitter andreceiver have a clear line-of-sight path. The received power at receiver P ris given bythe Friis free space equation [10]:2P G G λt t rP ( d)=r 2 2(4π) d Lwhere P tis the transmitted signal power, G tand G rare the antenna gains of the transmitterand the receiver respectively. L (L≥1) is the system loss, and λ is the wavelength.The two-ray ground reflection model considers both the direct path and aground reflection path between the transmitter and receiver. At a long distance, thismodel is more accurate than the free space model. The received signal power at adistance d from the transmitter can be expressed by [10]:P ( d)r=2PG G h ht t r t r4dwhere h rand h tare the heights of the transmit and receive antenna respectively.The free space model and two-ray ground model predict that received power decaysas a function of distance, the radio transmission range is a perfect circle. The shadowingmodel is a statistical model. The mean received power at distance d is computedrelative to P r(d 0), represented by (in dB) [10]:P ( d)[ ]P ( d )d= −10β log( ) X σdr+dBr 00Equation (3) is also called log-normal shadowing model. In this paper, we will simplyrefer to it as shadowing model, and the free space and two-ray ground model as idealmodel or ideal environment. The shadowing model consists of two parts. The first oneis the path loss model, d 0is a reference distance, β is called loss exponent. The secondpart reflects the variation of the received power at certain distance. X σ is a Gaussian2(1)(2)(3)