Stable adaptive control for nonlinear systems using neural networks

Stability analysis of neural-network-based nonlinear control has presented great difficulties. For a class of affine nonlinear systems with uncertainties, we employed nonlinear-parameter-neural-networks(NPNN) to approximate on-line the unknown nonlinearities, estimate on-line the NPNN approximation error's bound, and then succeeded in designing the control law and the adaptive laws of NPNN's weights and the NPNN approximation error's bound. The stability of the closed-loop is proved by using Lyapunov theory. Simulation results show that the controller we proposed exhibits excellent tracking performance.