关键词: 风管清扫机器人, 障碍物检测, 光流, 人工势场

Abstract: Duct cleaning robot is a kind of mobile manipulator working in abominable environment with complicated spatial disorder to clean the inner wall of central air-conditioning ventilation pipe. The collision-free movement of robot is crucial for the realization of its 3D non-repeating touch and optimal traversal cleaning with high-speed rotating brush. In this paper, the optical density is used as a source of obstacle avoidance. Combining optical flow and artificial potential field method, obstacle detection based on optical flow method, the artificial potential field method for obstacle avoidance planning, it proposes a local path planning for obstacle detection and avoidance. Experiment shows that repulsive potential field vector points to the direction of avoiding obstacles, and the potential field is mainly affected by the optical flow of obstacles and determined by the nearest obstacles. By contrast with the balance strategy, simulation results verify the effectiveness of this method.

Key words: duct cleaning robot, obstacle detecting, optical flow, artificial potential field