Page - 12 - in Maximum Tire-Road Friction Coefficient Estimation

1 Introduction application in an AEB. This system has been proven to have high potential benefit for avoiding and mitigating accidents, see Section 1.3.1. In addition, the requirements for pre-firing airbags are shown as an example for tertiary safety systems. Requirements for automated emergency braking systems Investigations have shown that many rear-end crashes happen because drivers do not brake hard enough, react too late or fail to react, [Bre09]. Braking assist systems (BA) assist drivers who do not brake hard enough by increasing the braking pressure when the braking pedal is actuated with a high velocity. In contrast, AEB systems come into play when drivers react too late or fail too react. The AEB intervention phase consists of two stages. In the first stage, a partial braking manoeuvre is initiated. The vehicle speed is reduced while drivers are still offered the chance to provide their own interven- tion. If thedriver fails to react, a full brakingmanoeuvre is initiated in the secondstage. The required accuracy depends on the relative velocity between the vehicles, the road condition and the braking system conditions. When the activation times of an AEB are adapted by an underestimated friction potential, the calculated braking distance is longer than the actual value and the intervention is started earlier. The collision is avoided, but valuable activation time is lost for collision validation, and driver accep- tanceof thesystemmaythereforedecrease. Whenthe frictionpotential isoverestimated, the collision cannot be avoided which results in an impact. Lex et al. determined that an impact speed of 4.2 m/s is tolerable in terms of pedestrian safety and vehicle re- pairability, [LKE13a]. Figure 1.8 shows the requirements for an estimate of the friction potential to be applied in an AEB when approaching a standstill object. The collision model uses Newton’s equation of motion for a single mass point with one single degree of freedom in the longitudinal direction. Such simple models are often used in published AEB warning algorithms, [Eic11, p.120]. Section 6.5 discusses the estimation algorithm for the friction potential developed in this thesis in the context of requirements for an AEB that are described here. For this algorithm, another representation of these requirements is used, which shows the tolerable deviation ∆µ versus the real valueµref; see Figure 6.16. 12