Fuzzy agents for reactive navigation of a mobile robot

The authors propose a sensor-based navigation algorithm built thanks to the fusion of various elementary behaviors. The proposed navigator combines two types of obstacle avoidance behavior, one for convex obstacles and one for concave ones. To avoid convex obstacles the navigator uses either a fuzzy tuned artificial potential field (FTAPF) method or a behavioral agent. The concave obstacle avoidance behavior results of “wall-following” behavior combined with the creation of transition subgoals. An automatically online tuned fuzzy wall-following system using a neuro-fuzzy structure is designed. The incorporation in the learning cost function of a weight decay term prevents an excessive growth of the weights and allows quick and efficient learning leading to a robust controller optimized with respect to the actual physical characteristics of the robot. The effectiveness of the proposed method is verified by carrying out experiments on the miniature mobile robot Khepera^(R)