Neural network control for nonlinear systems

A neural network controller is constructed for robust asymptotic set-point tracking in a class of nonlinear systems. By training the neural networks using the proposed algorithm, the set-point tracking in nonlinear systems and the convergence of the neural networks can be achieved. The convergence of the system is shown to be governed by not only the plant characteristics but also the initial conditions of the plant and controller. Simulation results show that the convergence of the system can be guaranteed by selecting the proper initial conditions of the plant and the neural network controller and the appropriate updating rate of the weights of the networks