Optimization-based motion primitive automata for autonomous driving

One of the challenges before the introduction of automatic driving are unsignalized inner city intersections. There deadlock situations are possible, which are not cleary regulated. Those involved have to cooperate in that case, which is challenging for automatic vehicles. The presented algorithm is able to make behavior decisions for automatic vehicle guidance. This decision depends on the situation and the behavior of the remaining traffic participants. The decision-making model is designed as a discrete event system, which is especially beneficial regarding the transparency of the decisions. The model is presented in detail alongside the features and events used for it and is validated by a simulation on maps of real intersections.