On a Laser Rangefinder-based Mobile Robot Obstacle Avoidance Strategy via Local Virtual Target

A reactive obstacle avoidance strategy based on local virtual target following is proposed in this paper with reference to obstacle avoidance problems for a unicycle mobile robot in an unknown environment. The strategy took a laser rangefinder as the main environment detecting sensor. A polar co-ordinate vector map (PMap) was utilized to store environmental information and a cluster ideology was used to describe obstacle features. The local virtual target comprised environment features and target information together and so it was used as a temporary target to guide the robot to track the real goal while accomplishing obstacle avoidance. This method could be applied to obstacles with arbitrary shapes and could reduce the environment memory amount dramatically. It improved the dynamic response performance of the robot and fulfilled real time obstacle avoidance when the robot was following a target. A great deal of robot experiments showed that the propounded strategy was very effective.