Approximation of a neural network controller based on model reference technique to identify a non-linear system

In this paper, a neural network (NN) controller is approximated by method of `training´, using a reference model, to identify a plant. At first, the plant identification neural network, is realized by learning the behaviour of a given nonlinear plant. The outputs of this neural network and the actual plant are compared. The knowledge of the comparison is fed back to the NN controller as its input. The NN controller is realized in such a way that it will control the plant behaviour as a reference.