Obstacle avoidance based on scene matching for mobile robot

Abstract

The problem of obstacle avoidance for mobile robot in unknown environment is studied. A new obstacle avoidance method based on fuzzy scene matching is presented. Using sensor fusion, scene corresponding to current environment, which matches with the scenes in scene base, is obtained. According to the matching result, robot′s motion parameters can be achieved through a fuzzy controller. These parameters are used to control robot for (obstacle) avoidance. Experiment results show the effectiveness and correctives of the method.