A hybrid neuro-fuzzy system for sensor based robot navigation in unknown environments

This paper presents a hybrid neuro-fuzzy system for sensor based robot navigation in unknown environments. A neural network is used to process range information for determining a good reference motion direction in local regions; while fuzzy sets and fuzzy rules are used to formulate reactive behavior quantitatively and to coordinate conflicts and competition among multiple types of behavior efficiently. This neuro-fuzzy system is used to control the THMR-II mobile robot that is equipped with an array of ultrasonic sensors to acquire distances between the robot and obstacles. On the basis of this system, the author proposes a strategy for combining low-level behavior control with high-level global geometric planning.