Behavior architecture fuzzy controller for unicycle robot

Obstacle avoidance, researching goal and the path planning are the most important tasks for autonomous mobile robot which moves in an uncertain environment, this paper presents a new method of fuzzy behavior-based controller in order to control the robot to reach the goal and avoiding all fixed and mobile obstacles. The developed control is based on three fuzzy controllers; the first allows the robot to reach the goal by calculating the shortest possible path and reducing the computation time, the second controller based on a new basis of fuzzy rules which allows the mobile robot obstacle avoidance better, the third controller is a controller fusion allowing the activation of one of the first two controllers according to the state of the robot, switching between the two controllers is a partially in order to give better stability in the movement of the robot, we use the Mamdani inference system to construct all controllers, and the centroid of area method as our defuzzification strategy where the inputs are considered as linguistic variables. Simulation results are obtained to demonstrate the feasibility of the implementation of this new method for autonomous mobile robots.