Sliding mode control based on self-organized fuzzy neural networks for multi-link robots

A sliding mode controller based on self-organized fuzzy neural networks (SMCSOFNN) is presented for trajectory tracking control of multi-link robots with model errors and uncertain disturbances. This approach gives a new global sliding mode manifold for the second-order robot manipulators, which enable system trajectory to run on the sliding mode manifold at the start point and eliminate the reaching phase of the conventional sliding mode control. Robustness for system dynamics is guaranteed over all the response time. Self-organized fuzzy neural networks (SOFNN) are employed to eliminate chattering of global sliding mode control, and enforce the sliding mode motion by SOFNN learning the upper bound of model errors and uncertain disturbances. SOFNN can optimize fuzzy rules according to the controlled system real-time accuracy. Therefore, the controlled system accuracy is improved. The control law and the cost function of the fuzzy neural network are calculated by Lyapunov stability method. Simulation results verify the validity of the control scheme.