A simple rule based multi-agent control algorithm and its implementation using autonomous mobile robots

Recently, many researchers have paid attention to researches on flock behavior in multi-robot system. In most of those researches, an absolute coordinate system is used for identifying a position of each agent consists a flock. It is costly to obtain the position in the absolute coordinates under practical applications because special devices, such as GPS, are necessary. Therefore we have developed an effective simple rule based multi-agent control algorithm for practical uses, which uses limited local information surrounding an agent instead of the absolute coordinates to obtain the situations. In this paper, we implement the algorithm into an autonomous mobile robot. The performance of the proposed algorithm was investigated by experiments using autonomous mobile robots having the proposed algorism. Three situations were examined: 1) guided running according to an external operator's command, 2) obstacle avoidance, and 3) adaptive cruse by changing a flock formation according to a bottleneck path. In all situations, robots move maintaining a flock form. The validity of the proposed algorithm is discussed through the experimental results.