Adaptive terminal sliding mode control for AC servo system based on neural network

Terminal sliding mode variable structure control (TSMC) is a special nonlinear control strategy, which has strong robustness against parameter variations, load disturbances and uncertainty of system. Nevertheless, the control precision and stability of system will be affected by the chattering, caused by sliding mode switch control. This control method can approach the nonlinear model through RBF neural network when the accurate model of the nonlinear system is unknown. The parameter adaptive law is designed according to Lyapunov theory, without sacrificing the strong robustness of TSMC. Simulation results show the control input is chattering-free, and both load disturbance and parameter perturbation can be compensated.