Centre of mass avoidance planner using radius of gyration for Reciprocal Velocity Obstacles

In this paper we presented a planning approach for collision avoidance algorithms in swarms of robots. We used intermediate goal vector formation to avoid congestion of large number of robots trying to maximize their performance measure. We use the RVO2 library which uses the Reciprocal Velocity Obstacles approach for collision avoidance in intelligent swarms. RVO is an excellent collision avoidance algorithm that can successfully solve an intelligent swarm situation. We have tried to improve its applicability and smooth operation in situations with eminent possibility of traffic congestion. With our approach, before finally reaching the intended goal, the robot follows an intermediate local goal to optimize the path of all robots by avoiding the imaginary centre of mass of the swarm agents as well as proceeding to the final goal. The centre of mass avoidance is triggered by the radius of gyration which signifies the current magnitude of congestion. We have also proven that our strategy is generic in nature and scalable to any number of robots. We also prove the computational ease of deploying the planning approach.