Distributed receding horizon planning for multi-robot systems

Abstract

This paper considers the navigation of a group of nonholonomic autonomous mobile robots subject to constraints imposed by sensors. The proposed scheme uses a decentralized receding horizon motion planner, in which each robot computes its own planned trajectory. The proposed scheme is divided into two steps. First, each robot computes its presumed trajectory without taking the coupling constraints (i.e. the collision avoidance between robots) into account. Then, robots which may produce collision adjust their presumed trajectory using local exchanged information. The proposed algorithm, based on potential field method, allows the navigation of a large-scale multi-robot system. Finally, simulation results demonstrate the effectiveness of the proposed strategy.