Virtual pheromone map building and a utilization method for a multi-purpose swarm robot system

Abstract

Swarm robotics is an approach to realizing an intelligent system using swarm intelligence that emerges from complex interactions among individual robots. To implement a swarm robot system requires an efficient method for sharing and utilizing the information that is gathered by multiple robots. In this paper, we propose an algorithm to gather, simplify, share, and use the information in a multi-purpose swarm robot system. To minimize communications traffic among the robots, we use a graph-based map-building method. Furthermore, the nodes on the graph include information about the environment for control and interaction among the robots. We refer to this graph-based map as a virtual pheromone map. We propose an efficient algorithm to build a virtual pheromone map with low communications traffic and without contradictions using error-included sensors. The utilization method of the virtual pheromone map is efficient for implementing the interactions among the robots. The proposed algorithms are validated by a computer simulation. The result of the simulation shows that the virtual pheromone map-based swarm robot system works efficiently from the effects of swarm intelligence.