Model-based recurrent neural network for redundancy resolution of manipulator with remote centre of motion constraints

Redundancy resolution is a critical issue to achieve accurate kinematic control for manipulators. End-effectors of manipulators can track desired paths well with suitable resolved joint variables. In some manipulation applications such as selecting insertion paths to thrill through a set of points, it requires the distal link of a manipulator to translate along such fixed point and then perform manipulation tasks. The point is known as remote centre of motion (RCM) to constrain motion planning and kinematic control of manipulators. Together with its end-effector finishing path tracking tasks, the redundancy resolution of a manipulators has to maintain RCM to produce reliable resolved joint angles. However, current existing redundancy resolution schemes on manipulators based on recurrent neural networks (RNNs) mainly are focusing on unrestricted motion without RCM constraints considered. In this paper, an RNN-based approach is proposed to solve the redundancy resolution issue with RCM constraints, developing a new general dynamic optimisation formulation containing the RCM constraints. Theoretical analysis shows the theoretical derivation and convergence of the proposed RNN for redundancy resolution of manipulators with RCM constraints. Simulation results further demonstrate the efficiency of the proposed method in end-effector path tracking control under RCM constraints based on an industrial redundant manipulator system.