In this paper, the realization of cooperative knowledge for a catching problem of multiple hunter robots and an escaping target is considered. To achieve successful catching, a pincer attack movement between two hunter robots is introduced as the basic cooperation strategy. The pincer attack movement of each hunter robot is described by the fuzzy rules. The fuzzy rules are tuned by genetic algorithm that the hunter robots can realize the pincer attack formation. Furthermore, the indirect approach is discussed in which the tuned fuzzy pincer attack rules are only applied to the nearest pair among the multiple hunter robots. Experimental and simulation results demonstrate that the suitable cooperative movement among the robots is realized and the proposed cooperation scheme is effective to the catching problem.
J-STAGEがリニューアルされました! https://www.jstage.jst.go.jp/browse/-char/ja/
