A hierarchical fuzzy controller for real-time mobile robot navigation in unknown environments

This paper presents a hierarchical fuzzy controller for mobile robot navigating from an initial position to a target point through unknown environments. Eight sonar sensors are mounted on the robot to detect obstacles so that the mobile robot can navigate safely. Sonar data and target information are processed by the two-stage fuzzy system to generate the control commands. Firstly, the danger fuzzy controller (DFC) is used for each sonar sensor to judge the danger degree of the detected obstacles. Then, all the outputs of the inference engine are analyzed and normalized by the combination behavior. Finally, the danger information as well as the fuzzified target information is induced by the velocity fuzzy controller (VFC) and the rotation fuzzy controller (RFC) respectively to determine the motion of the robot. Maximum defuzzification method is used to solve the "symmetric hesitation" problem. Simulations and experiments validate the performance and effectiveness of the proposed approach.