Cooperative and Intrinsically-Correct Control of Vehicles in Diverse Environments (CoInCiDE)


Automated driving will unquestionably provide a variety of benefits, including the reduction of traffic accidents, traffic jams, and mobility barriers. It is unanimously agreed among experts that the benefits of automated driving can only be fully used when vehicles cooperatively adapt to each other. A major obstacle towards this promising future is to guarantee in a mathematically provable way that the automated system always works correctly in diverse environments with varying road and weather conditions and a varying degree of collaboration of traffic participants. This endeavor is especially challenging since each traffic situation is different and thus automated driving cannot be verified beforehand, requiring that the vehicle verifies itself periodically given the current traffic situation. We address the important requirement to guarantee safe and coordinated motion of automated vehicles by establishing a framework, which integrates vehicle behavior prediction, cooperative maneuver- and trajectory-planning, and formal verification in a unique framework to provide intrinsic correctness.



  • On-the-fly consideration of traffic participants: The architecture should make it possible to add or remove traffic participants into the planning and safety analysis of an automated vehicle.
  • Unified approach for mixed traffic: The agent-based architecture should make it possible to easily integrate automated and manually-driven vehicles.
  • Consideration of uncertainties: All conceivable uncertainties of the vehicle motion within given bounds have to be considered. Furthermore, the planning horizon should enable the cooperative vehicles to compensate for short and intermittent imperfections of communication.
  • Intrinsic safety: Any computation can be stopped anytime and the cooperative group has to be able to perform a maneuver until a safe stop is reached.
  • Anytime computations: The planning of driving strategies and the refinement of driving strategies should follow an anytime principle.





  • Stefanie Manzinger, M.Sc.
  • Edmond Irani Liu, M.Sc.
  • Prof. Dr.-Ing. Matthias Althoff