Dynamic Trajectory Planning Of Collaborative Robots In Target Capture Task

Aiming at the trajectory planning problem of robots in collaborative surroundings, especially for dynamic obstacles, like human body, a real-time collision avoidance approach is proposed. Firstly, RGBD camera Kinect is utilized to recognize and locate the human body. And then, the distances of human-robot are obtained through a new sample method. Moreover, the trajectory is generated by the proposed approach(APF-RV), which combines artificial potential field method(APF) with repulsive vector method(RV). In order to improve safety and efficiency of robots, this method does not only consider static obstacles, but also takes into account the movement of dynamic obstacles. Then, simulation experiment is utilized to execute the new method and optimize parameters. Besides, further physical experiment is conducted by UR5 robot to verify the feasibility and effectivity of the designed real-time collision avoidance algorithm.