Kinematic Control for Redundant Manipulators with Remote Center of Motion Constraint based on Neural Network

In this paper, a dynamic neural network-based kinematic controller is proposed for redundant manipulators under the remote center of motion (RCM). The constraint of RCM is built into an equality constraint and is then written as an augmented equality constraint together with the end-trajectory tracking mechanism. Considering physical limitations such as joint angles and velocities, the kinematic control problem is established as a dynamic programming system, and then a dynamic neural network-based controller is built to obtain the control command recurrently in real-time. Numerical experiments show the effectiveness of the proposed controller in both static and dynamic RCM situations.