Distributed multi-robot formation control under dynamic obstacle interference

When multi-robot systems are keeping formations in dynamic environments, dynamic obstacles may occasionally enter the formations. Under this interference, the sight of robots may be blocked and the collision may occur, thus the topology of the formation may also change. Inspired by shoals of fish fleeing dolphin predation, a novel distributed formation control method based on the artificial potential field (APF) and distributed consensus has been proposed. For dynamic obstacles, we extend the velocity potential field in the velocity direction of the obstacles. By combining the formation speed and the collision avoidance speed, the flexible switching of the formation maintenance and obstacle avoidance can be realized. The experiment results show that the teams can maintain the original formation as much as possible or they can divide into sub-formations to avoid collision and finally reconvene.

I. Introduction

Multi-robot systems have been increasingly employed in research and industrial environments in recent years. As an important research direction of multi-robot system, formation control can be applied in multiple scenarios, such as searching or surveilling an area, maintaining communication network, collaboratively handling an object [1]. Many methods of formation control have been proposed, including behavior based control [2], artificial potential field based control [3] [4], virtual structure approach [5], distributed consensus algorithms [6], leader and follower approach [7], and graph based formation control [8]. However, most of these studies consider an obstacle-free environment, whose focus is mainly on preventing collisions between robots.