A Neural Network-Based Adaptive Dynamic Surface Control for Nonlinear Systems in Strict-Feedback Form with Input Constraints

In this paper, an adaptive tracking controller that takes into account actuator saturation is designed for a nonlinear system with arbitrary uncertainty. In traditional backstepping design, the virtual controller at each step requires repeated differentiation, which leads to design complexity explosion as the system order increases. By introducing a first-order filter and using dynamic surface control, the problem of design complexity explosion associated with traditional methods is solved. In addition, the RBF neural networks are used to solve the approximation problem of nonlinear functions. At the same time, in order to solve the common input saturation constraint problem in engineering, an auxiliary error compensation system is introduced to enable the controller to better cope with input saturation. Finally, Lyapunov stability analysis shows that the proposed control method ensures the stability of the closed-loop system and reduces the tracking error to a negligible level asymptotically. By introducing an anti-saturation auxiliary error compensation system in the controller, the system can be stabilized when input saturation occurs.