Accuracy preserving interpretability by a hierarchical fuzzy multiagent design for mobile robot controller

Abstract

This paper proposes an innovative multiagent design for fuzzy controller of a multisensory Khepera II mobile robot. Undeniably, the reactions' speed, the precision in motion and how easy to understand these reactions are the main criteria for such a fuzzy controller. Indeed, an agentification of sensors and distribution of control over several cooperative fuzzy-agents are the strategic points for the proposed model. An implementation followed by a series of simulations of the Khepera navigational behavior, allowed us to test and discuss the efficiency of the fuzzy-multiagent design on the accuracy preserving the interpretability for the problem of obstacle avoidance in unknown and dynamic environment.