Autonomous Vehicles & Vessels

Autonomous driving is among the most media-present scientific topics today. Autonomous road vehicles are envisioned to revolutionize traffic in various ways, e.g., by increasing road safety due to fewer accidents and better traffic flow, or by decreasing vehicle ownership due to car sharing. Similar motivation propels research for autonomous vessels, as most collisions at sea happen due to human errors. However, the fundamental question of how to generate a safe trajectory, such that a vehicle or vessel reaches its destination despite unforeseeable circumstances still remains unanswered.

Research questions include

• How do we rigorously formalize traffic rules?
• How can we predict the behavior of other traffic participants?
• How can we plan a drivable trajectory despite uncertain disturbances?
• How do we ensure that our planned trajectory respects traffic rules?
• How can we efficiently re-plan a trajectory if the circumstances change?
• How can we leverage reinforcement learning to make high-level decisions?
• How can we safe-guard learning algorithms to guarantee safe trajectories?
• How can we optimally represent the road network for an efficient trajectory planning algorithm?

Our research on autonomous vehicles and vessels is implemented within the open-source software frameworks CommonRoad and CommonOcean, respectively.