A Radial Basis Function Neural Network Model Reference Adaptive Controller for Nonlinear Systems

In this paper, an algorithm of a model reference adaptive controller for nonlinear systems based on Radial Basis Function (RBF) Neural Networks (NN) is proposed. An adaptive RBF NN is trained using the error between the system response and the desired response as given by the RBF NN reference model. The stable controller-parameters adjustment mechanism is based on the convergence of the synaptic weights, centers and widths of the RBF network. Tuning the parameters of the RBF neural network is performed such that control error reduction and improved tracking accuracy are accomplished. The control signal obtained can make the real system close to the reference model. The identification and control error, with acceptable convergence speed, have been achieved in the case of third-order non-linearity.