Multi-robot search using a physically-embedded Particle Swarm Optimization

We have derived a version of the PSO algorithm that is suitable for a swarm of small, mobile robots. Our version of the PSO is physically embedded into the robots; there is not a central agent that coordinates the movements or actions of the robots. Simulation results show that the algorithm is very good at finding a target location within a complicated search space and is scalable to large swarm sizes. We built three robots and tested the new algorithm by letting the bots find the brightest spot of light in the room. All multi-bot searches were successful with the new algorithm and were faster than non-PSO searches. We have demonstrated a successful implementation of PSO into a hardware swarm of robots.