Page - 107 - in Maximum Tire-Road Friction Coefficient Estimation

6. Results and conclusion This chapter presents the results for several driving manoeuvres, including accelerating andbraking. The identificationof the frictionpotentialonhighand low-frictionsurfaces, as well as for changing road conditions (µ step) and different road conditions on the left and right wheels (µ split) were investigated. In addition, examples of the influence of the longitudinal acceleration on the estimation accuracy is shown. Concerning the particlefilter, this chapterdiscusseshowtheevolutionofparticlesversus timeaffects the µmax estimate and shows strategies for influencing the course of particles. The achieved accuracy of the friction potential estimate is discussed with respect to the requirements established in Section 1.3.2 for the application in an autonomous emergency braking system (AEB), which was chosen as an exemplary application of a friction potential estimate to adapt an ADAS activation strategy. This chapter then concludes with a general discussion of the results and provides an outlook. 6.1. Strong acceleration manoeuvre with constantµmax Figure 6.1 shows different hypotheses ofFx,i (gray) for different particles ofµ max during an acceleration manoeuvre. As described in Section 5.2,N hypotheses of longitudinal tire forcesFx,i are calculated for allN particles ofµ max for the inputsof the longitudinal slip sx, the side slip angleα and the tire loadFz,i of the current time step k. These are then compared to the expected tire force (black) at each time step, which is calculated based on the engine torqueMD,i or with the vehicle’s longitudinal acceleration bax, cf. Section 5.5.3. It canbe seen inFigure6.1 thatbetweenabout second4.7and5,Fx,i(MD,i) (black) is momentarilyhigher than ispositedbyanyof thehypothesisofFx(particles) (gray). This can happen due to inaccuracies in the calculation of either Fx,i(MD,i) or Fx,i(sx,i,N). In this case, the highest particle is assigned the highest probability. Although it is not possible to extrapolate to conditions outside of the minimum and maximum particle values, interpolation between the discrete particle states is carried out at each time step.