The introduction of autonomous cars in street traffic creates the desire for self-organization of these cars regarding security, economic and ecologic aspects. To this regard, optimization based distributed control algorithms provide a solution approach for communicating autonomous cars. In this work we discuss a hierarchical model predictive control concept. On the operational level, individual goals of the drivers and security aspects are considered. Utilizing a planning layer, we show how overall goals like traffic flow maximization can be implemented. To this end, if a traffic event is triggered, a route for each individual car is computed centrally based on the current traffic conditions. The structure of the control also allows integrating cars, which are not actively taking part in the communication, via sensor fusion techniques.
