Neural-networks-based adaptive asymptotic tracking control for nonlinear systems with periodic disturbances

A dynamic surface control scheme based on neural networks is proposed for a class of uncertain nonlinear systems with periodic disturbances and input delay. Combining the radial basis function neural network (RBFNN) with the Fourier series expansion (FSE), a mixed function approximator is constructed to approximate the unknown periodic disturbances functions in the system. An integral term is introduced to solve the problem of input delay. At the same time, the dynamic surface control method with the nonlinear filter is developed to avoid the problem of explosion of complexity commonly existed in the adaptive backstepping control method. The semiglobal boundedness of all closed-loop signals is guaranteed, and the output of the system can track a given reference signal asymptotically. Finally, two simulation results show that the proposed control scheme is effective. Copyright ©2022 Control and Decision.