Adaptive robust dynamic surface control of pure-feedback systems using self-constructing neural networks

This paper proposes an adaptive robust dynamic surface control (ARDSC) method integrated a novel self-constructing neural network (SCNN) for a class of complete non-affine pure-feedback systems with disturbances. By employing the mean-value theorem and implicit function theorem, the adaptive robust control (ARC) method is extended to pure-feedback systems, and improves the robustness and transient performance of the closed-loop system. The "explosion of complexity" in backstepping scheme is avoided via dynamic surface control (DSC) technique. Moreover, the controller complexity is further reduced by introducing an SCNN based on a novel pruning strategy and a width adjustment strategy. Input-to-state stability and small-gain theorem are utilized to analyze the stability of the closed-loop system. At the end, simulation results demonstrate effectiveness and advantages of the proposed control method.