A real-life robotic application of the particle swarm optimization algorithm

Abstract

This paper describes our effort to use the well-known particle swarm optimization algorithm to control a small swarm of robots in a search and rescue application. The algorithm was slightly modified from its original format to better fit real-life robotic applications. The paper discusses the modified algorithm and explains the challenges that are faced during a real-life implementation as opposed to the traditional, simulation-based, virtual implementation. The paper also describes the experimental setup and the robots used in the experiment. When implemented on the robotic swarm, the experimental results show that the modified algorithm was consistently able to find the global maximum point in no more than 80 seconds. The results also showed that changing the robots' increment size (between 9.75cm and 16.25cm) did not have a significant effect on the efficiency of the swarm or its convergence.