Adaptive Neural Network Based Virtual Decomposition Control for Robot Manipulator with Mixed Rigid/Flexible Joints

Abstract

In this article, an adaptive neural network based controller for multi-degree-of-freedom robot manipulators with mixed rigid/flexible joints is investigated within frame of virtual decomposition control theory. First, virtual decomposition principle is introduced and applied to decouple the entire system of robot manipulator with mixed rigid/flexible joints into subsystems in terms of links and joints. Then, generalized neural networks are incorporated into the control law of rigid-link subsystems and rigid-joint subsystems to make the corresponding sub-control without invoing any model parameters. In addition, overall stability analysis of the control system is given according to the theory of virtual stability and Lyapunov stability. Finally, validation example is provided to verify the proposed control approach.