13.07.2015 Views

Advance Modeling of a Skid-Steering Mobile Robot for Remote ...

Advance Modeling of a Skid-Steering Mobile Robot for Remote ...

Advance Modeling of a Skid-Steering Mobile Robot for Remote ...

SHOW MORE
SHOW LESS
  • No tags were found...

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

6.2 Simulation Results 876.2 Simulation ResultsIn this section the results <strong>of</strong> the simulation are presented and further discussed to characterizethe properties <strong>of</strong> the model, and they are compared with the real data acquired withthe accelerometers in order to check the model reliability. Be<strong>for</strong>e running the simulation, weneed to estimate the values <strong>for</strong> the damping coefficients D x , D y , D z . As explained in Section5.2, it was not possible to analytically compute the values <strong>for</strong> D x , D y , D z from the identifiedvalues <strong>of</strong> the angular natural frequency and the damping ratio <strong>for</strong> the Roll, Pitch and Yawmotion. Conversely, an estimation <strong>of</strong> the tire damping coefficients can be obtained by settingD x , D y , D z in the simulation to some random values and then changing those values so thatthe Roll and Pitch free responses match with the data presented in Section 5.2. We also recallthe fact that, as the third equation in (5.3) holds <strong>for</strong> the acquired Yaw free response, theparameters D x and D y are dependent from each other by the second relation on the right sidein (5.10). Thereby, we only need to set the values <strong>for</strong> D z and either D x or D y .The initial condition <strong>for</strong> the free response is provided by setting to different values the initialcondition <strong>of</strong> the integrator <strong>for</strong> the vertical reaction <strong>for</strong>ce. In particular, <strong>for</strong> the Rollmotion we set the same initial condition <strong>for</strong> the two left and right-side wheels, while <strong>for</strong>the Pitch angle we set the same condition <strong>for</strong> the two front and rear-side wheels. In orderto provide a free response as close as possible to the acquired one, the initial conditionsare set symmetrically with respect to the tire compression at equilibrium, that is <strong>for</strong> theRoll case z init1 = z init2 = − mg4Ktire z− 0.001 and z init3 = z init4 = − mg4Ktire z+ 0.001. As the responseis under-damped in the three cases (see Figure 5.4, 5.5), the damping coefficientsare initially set to one fourth <strong>of</strong> the value corresponding to a critically damped response, i.e.D x = 1 2√ mKx , D y = 1 2√ mKy , D z = 1 2√ mKz , and then they are either increased or decreaseddepending on the simulation results. After several trials, by comparing the results from the¨Accelerometers¨ block with Figure 5.4, 5.5, we obtained the following values:D x = 50, D y = 30, D z = 70 (6.5)Figure 6.5 and 6.6 depict the simulation results respectively <strong>for</strong> the Roll and Pitch motion,obtained from the ¨Accelerometers¨ block by setting the values in 6.5.It is worth noticing that, although this estimation is not accurate and could not be unique,it provides the needed in<strong>for</strong>mation to reproduce the real Roll and Pitch oscillation, which is

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!