Page - 10 - in Maximum Tire-Road Friction Coefficient Estimation

1 Introduction but one is detected anyway and the system intervenes, this false intervention can result in severe accidents. For this reason, all subsequent phases are activated as late as pos- sible, which leads to the so-called warning dilemma: with the validation time obtained, the probability of the predicted collision parameters in the pre-collision phase can be increased. But on the other hand, the driver warning is more effective the sooner it is initiated, [HG09]. Two factors are crucial for predicting a collision and therefore the activation times for the different phases, [Eic11, p.120-127]. These two factors describe the possible courses of the traffic participants involved in the pre-collision phase which are necessary to pre- dict the collision parameters. The first factor is the drivers’ reaction, e.g. whether there is operationof steeringwheel, brakesor throttle and, in caseof anoperation, howstrong the driver’s input is. The second factor is the friction potential that limits the maximum transmittable horizontal tire forces and therefore the possible motion of the vehicles in the horizontal plane. Depending on the warning strategy, the driver is being alerted to be either attentive or to set an action in the warning phase by acoustic, haptic or visual signals. The transition between warning and intervention is herein referred to as decision phase, as the course of action is chosen. Depending on the probability of the predictedcollision, an interventionhas tobeomitted (e.g. when the collisionprobability is low) or activated. Some systems also allow for different intervention strategies, such as braking or steering around an obstacle as shown in Figure 1.7. Whether it is optimal to brake or to steer in order to avoid a collision depends strongly on the friction po- tential. In addition, a system must take into account both the relative velocity and the possibility of steering around an object, which depends on the road type and other traf- fic participants. Finally, the intervention phase starts, provided that the driver has still notreacted. Thisphasecanconsistof several stages, seee.g. foranAEBinSection1.3.2. Current systems are developed to fulfill the requirements of collision mitigation and avoidance on dry roads. However, many ADAS show their highest potential for avoiding accidents or reducing the injury severity on low-friction surfaces. Thus, investigations have proven that the activation times of an EMA can be significantly increased on low-friction surfaces when the friction potential is known because interventions can be started earlier, [NLER11] and [LRNE13]. A combined consideration of driver behaviour and vehicle dynamics results in the following thresholds for activation times, [Win09a]. Whenanimminentcollision isdetected, thetheoretical timeuntil theobstacle is reached, 10