CityTrix: Constrained Motion Planning for Automated Complex Vehicle Combinations

We present a novel algorithm for motion planning for a special kind of robots: automated vehicles with complex geometry, such as trucks with trailers. The proposed CityTrix algorithm and its associated software architecture enable collision-free and kinematically feasible path planning for automated vehicles on a provided network of roads. The algorithm was developed within German and European research projects for vehicle automation in delimited areas such as logistics centers or maritime ports. Within this paper, we discuss the requirements on motion planning within these so-called yards. We present a software architecture to overcome major limitations of previous planners and show, how CityTrix can effectively calculate drivable paths for large vehicles in challenging scenarios. The main advantage of the proposed method lies in the adaptability to a wide range of scenarios due to using abstraction.