结合光流和人工势场的风管机器人避障方法

Obstacle detection method for duct cleaning robot combining optical flow and artificial potential field

风管清扫机器人是一种用来清扫中央空调通风管道内壁的机械手,工作环境恶劣、管道空间障碍复杂。风管清扫机器人的无碰撞运动对于实现其3D非重复接触和高速旋转刷的最优遍历清洗至关重要。以光流密度作为使用避障的光源,结合光流和人工势场法,以光流法进行障碍物检测,人工势场法进行避障规划,提出一种局部路径规划的障碍物检测及避障方法。实验结果表明,斥力势场矢量指向避开障碍物的方向,而且势场主要受障碍物的光流的影响,并由最近的障碍物来确定。通过与平衡策略方法对比,也验证了此方法的有效性。

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.