Optimization of a Self-organized Collective Motion in a Robotic Swarm

Abstract

A novel collective organization method is proposed in this paper to improve the performance of the former Active Elastic Sheet (AES) algorithm by applying the Particle Swarm Optimization technique. Replacing the manual parameters tuning of the AES model with an evolutionary-based method leads the swarm to remain stable meanwhile the agents make a perfect alignment exploiting less energy. The proposed algorithm utilizes a hybrid cost function including the alignment error, interaction force, and time to consider all the important criteria for perfect swarm behavior. The Monte-Carlo simulation evaluated the algorithm’s performance to establish its effectiveness in different situations.