Predicting the Effect of Driver Assistance via Simulation

Abstract

Developing assistance systems in the automotive domain involves several exploration and evaluation activities with potential users of the system. To replace the amount of human test subject involvement, executable models reproducing human behaviour are introduced. Together with models of the car, road, surrounding traffic and of course the assistance system, a complete representation of the assistance system in its application environment can then be constructed which may be used to predict effects of introducing the assistance without having to resort to experiments with humans. In this paper we present techniques concerned with the exploration of the behaviour spectrum of the combined models. We show how functionality and safety aspects of assisted driving can be evaluated in computer simulations already during early phases of the development process.