Adaptive robust control for a class of nonlinear uncertain system based on neural network

By employing the neural network with the properties of approximating any continuous function with arbitrary, the online adaptive laws for weight matrix of neural network and robust controller are synthesized for a class of nonlinear system with uncertainties, whose the bounded functions is unknown. Through centralizing the uncertainties of the system to be controlled, the structure of the controller is simplified. Under some simple conditions, it is proved that the proposed controller can ensure both the states of system controlled and the approximation errors uniformly ultimately bounded (UUB). Finally, the simulation of the system composed of two water tanks shows the validity of the method.