Handling uncertainty in context-aware driver assistance systems

In this paper we discuss types of imprecision that are important for a tactical driver assistance system during its reasoning process. Methods for handling imprecision are presented using a previously developed overtake assistant as a showcase. The ideas for handling imprecision are demonstrated using several examples, together with the application of different approaches for different types of uncertainty. All presented solutions rely on simple efficient mathematical models, allowing for easy integration into the reasoning process without performance loss, while still yielding natural and useful quality assessments for sensed traffic information.