Indirect Adaptive Neural Network Control Using Dynamic Surface Control

Based on dynamic surface control, a novel design scheme of adaptive neural network controller is proposed for a class of perturbed strict-feedback nonlinear systems with unknown virtual control gain functions in this paper. The approach utilizes the differentiation of the first-order filter to replace the quantity of the differentiation of the virtual control in determining the next virtual control at each step of recursion. As a result, the operation of differentiation can be replaced by simpler algebraic operation. Therefore, the problem of explosion of complexity in traditional backstepping design, which is caused by repeated differentiations of certain nonlinear functions such as virtual control, is overcome by introducing the first order filter. Moreover, the possible controller singularity in feedback linearization is avoided without projection algorithm. Using Lyapunov method, the closed-loop systems is shown to be semi-globally uniformly ultimately bounded, with tracking error converging to a small neighborhood of origin by appropriately choosing design constants.