Efficient learning control of uncertain nonlinear systems with input constraints: a disturbance observer-based neural network approach

To deal with the effects of the input saturation and time-varying input delay, this article presents a serial-parallel identifier-based composite neural network learning control for uncertain nonlinear systems subject to external disturbances. Based on the backstepping technique, a radial basis function network is adopted to identify the unknown term, where the neural network learning accuracy is studied by considering a modeling error. In addition, a compensation system is designed to cope with input delay and input saturation, simultaneously. Besides, the explosion of complexity is mitigated by employing the command-filtered control approach. To enhance the robust performance of the overall system, the proposed control structure is enriched by a disturbance observer. Therefore, new adaptive rules are constructed. The stability of the closed-loop system is ensured by the Lyapunov theorem. Simulation results clarify the efficiency of the proposed control algorithm.