A novel neural sliding mode control for multi-link robots

A novel neural sliding mode controller is presented for trajectory tracking control of multi-link robots with external disturbances and uncertain system parameter errors. This approach combines neural networks and global sliding mode control. It adopts a global sliding mode manifold which eliminates reaching mode phase of conventional sliding mode control and robustness exists over all the system process. A radius basis function (RBF) neural network is applied to learn the system parameter errors and external disturbances. So the control system can automatically track the robot parameters and disturbances, and reduces chattering of the controller. Prediction estimation for robot parameters and disturbances is not needed too. Moreover, the system stability is proved by Lyapunov principle. Simulation results verify the validity of the control scheme.