Neural network-based control of nonlinear dynamical systems

This paper investigates neural network controllers to achieve output tracking for a class of SISO nonlinear systems. A discrete-time control strategy is introduced by means of employing the neural networks for direct on-line estimation of the required feedforward control input. The parameters of the networks are updated according to a modified backpropagation algorithm. Under appropriate conditions the stability of the neural network control system is rigorously analysed.