Page - 117 - in Maximum Tire-Road Friction Coefficient Estimation

6 Results and conclusion 2.5 3 3.5 4 4.5 5 0 0.2 0.4 0.6 0.8 1 1.2 1.4 Time in s 2.5 3 3.5 4 4.5 5 0 0.2 0.4 0.6 0.8 1 1.2 1.4 Time in s Figure 6.9.: Convergence and re-initialisation of variable particles marked by gray scale using the standard deviation (SD) of the variable particles (top) and the difference ∆µˆmax between the estimates of fixed and variable particles (bottom). 6.3. Gentle braking manoeuvre on varyingµmax Since the proposed method for estimating µmax requires a change in the driving state, the value of longitudinal acceleration bax acting on the vehicle’s body has an influence on the estimation accuracy. The closer bax is to the physical limits partly set byµ max, the more accurately µmax can theoretically be estimated. Thus, high values of bax, as previously discussed, are assumed to generally provide better estimation results. Figure 6.10 shows the estimates ofµmax for a braking manoeuvre with a maximum deceleration of≈−2 m/s2 as shown in Figure 6.11 that was performed during aµ step manoeuvre, starting on dry road (µmax≈ 1) and proceeding to icy road (µmax≈ 0.3) at second 9. Due to the significant difference between the low bax and the high µ max, the dry road condition cannot be detected at all. It appears that the low dynamic excitation is being confused with a low µmax. Regardless of the method, only the low-friction surface can 117