Prediction of Road Traffic Participants Using Reachability Analysis

With the big goal to develop a collision avoidance system, this thesis provides an important contribution in road occupancy prediction of traffic participants using reachability analysis. For computing occupied sets a vehicle model and a road model are introduced. Further, the driver behavior is restricted in terms of road traffic regulations and unrealistic actions. In addition, an analytical investigation regarding the limitations of these predicted sets is given. As a result and in connection with a trajectory planner, based on rapidly exploring random trees(RRTs), it is possible to assess whether and ,if so, when planned emergency maneuvers are collision-free at the last possible point of time.