Obstacle Avoidance of Swarms Using Pinning Control

In swarm control tasks, local objectives, on each agent, can interfere with the group's collective objectives. For example, in an environment with obstacles, the motion of the whole group can be affected by local obstacle encounters (happening in a few agents). In this work, we investigate the navigation of swarms in the presence of obstacles. We propose a novel control strategy to avoid obstacles while reducing swarm fragmentation, i.e., limiting the division of the swarm into disconnected groups. We model the swarm as a network where each vehicle is topologically connected with the neighbours that are within the agent's sensing range. We actively monitor the agents' connections in order to identify the necessity of redesigning the network, splitting a larger group into groups with fewer agents. Also, we use a path planning algorithm to provide the trajectory to guide the agents to the final destination. At the end of this paper, we show the results of simulation trials to demonstrate the performance of our control strategy. Copyright (C) 2020 The Authors.