A hierarchical fuzzy steering controller for mobile robots

This paper describes a hierarchical fuzzy steering controller approach to local navigation of an autonomous sensor equipped mobile robots. Two goals are considered simultaneously: obstacle avoidance, so that the robot is piloted safely in an unknown environment; and reaching an target point. A two-step approach to build a fuzzy system is presented. In the first stage signals are processed by conventional methods to get an output vector indicating the direction and the closeness of obstacles around. Stage two is a basic task steering controller, that is based on fuzzy logic. The modular structure of the first stage allows modifications on the robot. So there is no need for a redesign of the controller's architecture, when the controller is used in different applications. Adaptation to a new task is done simply by adjusting the controller to other conditions. Then the robustness of this controller is demonstrated in simulations by means of erroneous signals of the sensor modules. Finally the controller is used in combination with another basic task controller for 'Reaching an Aiming Point'. In a higher level a controller ascertains priority of the controllers and manages the resulting steering commands.