A direct adaptive neural-network control of nonlinear systems

A direct adaptive neural-network control strategy for a class of nonlinear system is presented. The system considered is described by an unknown NARMA model and a feedforward neural network is used to learn the system. Taking the neural network as a model of the system, control signals are directly obtained by minimizing either the instant difference or the cumulative differences between a setpoint and output of the model. To accelerate learning and improve convergence the technique in generalized predictive control theory and the gradient descent rule are used in this paper. The effectiveness of the proposed control scheme is illustrated through simulations.