Predictive curve anticipation

Parallel to assistant systems reacting to the close vehicle environment, it is necessary to provide a system, which can anticipate the road dynamics like stopping points or curves with small radius [1]. Indeed the vehicle safety is not only performed by looking for the current instant, but also by monitoring the speed and getting the possibility to decrease it in advance by taking into account the vehicle dynamics. In this paper, predictive curve anticipation will be described from sensors point of view and vehicle dynamics computation point of view. Its main task is to look at the curve ahead the vehicle by using both telematic platform with GPS and camera. The maximal safe speed for each point ahead will be defined by computing the maximal lateral force acceptable for the tires. From this maximal speed, the maximal safe acceleration will be extracted depending on the distance and the braking rate of the vehicle. Unfortunately if a crossing section appears ahead the vehicle, some doubts about the way chosen will be legitimated, and a Bayesian approach will be used to estimate which road will be taken and modulate correctly the speed.