Page - 37 - in Maximum Tire-Road Friction Coefficient Estimation

2 Estimation of the friction potential of 32 published vehicle-dynamic-based approaches, [LEH11]. In a second step, the pre- selected methods were compared using simulation data. This second step also took into account two additional evaluation criteria response time and response threshold of the estimates. Since the application is similar to the one in the present work, these criteria apply here as well. However, some criteria conflict with each other, such as not using additional sensors and enabling detection ahead of the vehicle. In addition, low-rated criteria such as gravel detection and availability for either tire, axle or a global vehicle as well as in the longitudinal, lateral or combined direction are not investigated in de- tail. However, combined estimation per tire, although not mandatory, is favoured. The investigation in Lex et al. also showed that there was no proposed method that worked under all driving states, [LEH11]. Most of the investigated methods were optimized for a particular driving manoeuvre, such as pure braking or pure cornering. Since it is a prerequisite that the algorithm proposed in this work functions for all driving manoeu- vres, this criterion has been added. Since the criteria for estimation accuracy and time delay depend on the application (see Section 1.3), they are considered in the criterion applicability to ADAS. As one important condition, robustness is a prerequisite for all automotiveapplications. These requirements leadtothemethodpresented inChapter5. Analternativemethodtoevaluatetheperformanceofestimators forthe frictionpoten- tial was presented by Bruzelius et al., [BSY+10]. This approach can be used to evaluate ready algorithms and sensors using vehicle measurements on different road surfaces. 37