The algorithm of autonomous obstacle avoidance based on behavior fusion for mobile robot

When the mobile robot facing unknown or uncertain environment,the weight of each behavior changes along with the environment change,resulting in that the weight can not adjust online,so the mobile robot can not arrive target point successfully.An autonomous obstacle avoidance behavior fusion algorithm that the weight can adjust online is presented.Mamdani type fuzzy controller is used to design basic behaviors in autonomous obstacle avoidance process for mobile robot.At the same time,a new method of weight adjustment based on two-stage BP neural network information fusion is presented to realize each basic behavior fusion to determine the weight of each behavior based on to fuse each behavior.This algorithm can realize online adjustment of each basic behavior weight according to environmental change and can strengthen real-timing and robustness of the autonomous obstacle avoidance control.The experiment results indicate that the algorithm is effective and feasible.