Adaptive NN Dynamic Surface Control for a Class of Uncertain Pure-feedback Systems

The tracking control problem of a class of pure-feedback system with uncertainty is considered. Dynamic surface control (DSC) design technique is used to avoid the problem of "explosion of complexity" in the traditional backstepping algorithm and overcome the circular construction of controller from function approximator in the backstepping design. Combining backstepping technique and neural network theorem, a semi-global uniformly ultimately bounded controller based on the Lyapunov stability theory is given. A simulation example is presented to demonstrate the method.