Mobile Agent Based Obstacle Avoidance in Multi-robot Hunting

Abstract

Multi-robot hunting is a problem in which multiple robots cooperatively search for a target that emits signals in the all directions. Robots proceed in the guidance of the signals, and finally, some of them reach the target to capture. In order to solve the problem, an approach based on Particle Swarm Optimization (PSO), which is one of meta-heuristics, has been proposed. The PSO based approach is well known that it works well in fields with no obstacle. It is, however, not assumed to be used in practical situations with obstacles. In order to lift this restriction, we propose a new PSO based approach that enables particles search and capture the target while getting around the obstacles. In our approach, each robot records its moving trace in a fixed period. Once a robot is blocked by obstacles and cannot proceed, it creates a mobile software agent that migrates to other robots around it through Wi-Fi. The mobile software agent selectively migrates to some robots whose traces have some intersection points. Since a sequence of the traces gives a detour route through which a robot can go, the agent can just inform the detour route to its home robot. We have implemented a simulator based on our approach, and conducted experiments. The experimental results show that our approach is remarkably more effective than the original PSO based approach.