Prioritized Multi-Robot Velocity Planning for Trajectory Coordination of Arbitrarily Complex Vehicle Structures

Abstract

Prioritized planning is a widely used technique to coordinate motions of multiple robots along fixed paths. Velocity profiles are planned sequentially avoiding previously computed trajectories represented as dynamic obstacles. To derive these, the vehicle structures are usually approximated with simple shapes, which fails for complex robots like automated truck-trailer combinations. A novel method to derive the spatiotemporal conflict zones directly from arbitrarily complex geometries swept along pre-planned paths is proposed and it is shown that the resulting obstacles in the distance-time-space allow velocity planning with precise space utilization and short execution times.