A fuzzy control system for autonomous navigation of mobile robots has been designed and implemented using behavior-based approach. We propose a novel concept of direction weights for sonar scanning of the robot environment. This concept facilitates navigation design using simple fuzzy logic controllers. Only three fuzzy rules are used in the fuzzy logic navigation controller. Thus the mobile robot has fast response to obstacles in an unknown environment. Fuzzy-Kohonen clustering network (FKCN) is adopted for sonar direction weights determination. In control implementation, the direction weights are adjusted in real time to the environmental changes. Both computer simulations and practical experimental results are presented to verify the effectiveness of the proposed navigation system.
展开▼
