Fuzzy logic navigation in multi agent systems

Abstract

In this paper we present a methodology for decision making in multi agent system comprised of agents driven by repulsive force and attracting force. Navigation functions are expressed as a set of fuzzy rules obtained by fuzzy Lyapunov stability criteria, thus ensuring stability of the overall system. The goal of the proposed methodology is to create desired formations by moving agents from their initial positions to formation targets, while in the same time avoiding collisions. The destination targets are dynamically permutated as long as the required formation is achieved.