Nonlinear systems identification using RBF neural networks

Abstract

We present a recursive nonlinear identification technique based on feedforward neural networks. A distinct feature of the proposed technique is the use of Radial-Basis-Function (RBF) neural nets as generic discrete nonlinear model structure. RBF nets have enabled us to devise a stable weight updating algorithm that guarantees the convergence of the weights to the target values. A simulation example is provided to illustrate the effectiveness of the method.