2A1-U06 Vision based model predictive control for the obstacle avoidance(Cooperation between Human and Machine)

Abstract

This paper proposes a control method for the task space obstacle avoidance problem of the robot manipulator. Proposed method utilized a vision based model predictive controller (MPC) in order to consider about structure and task space condition. In the method, current position of the obstacle is visually obtained and desired trajectory of the MPC controller will be updated to avoid the detected obstacle position. Details of the vision based MPC controller is mathematically described Experimental system using parallelogram linkage manipulator and Video-Tracker (the motion capture system) is developed to examine the proposed method. Experimental results show the effect of the method.