A model prediction for non-rational models via Radial Basis Function network

Rational models have been gradually adopted in various applications of nonlinear systems in the area of the systems identification and control because they have the advantage of modeling certain types of discontinuous functions and even severe non-linearilities using only a very few parameters. Based on the principle of Radial Basis Function even with noisy regressors, this work presents an alternative approach for the model prediciton of non-linear rational models. Two examples are included to show the performance of the proposed methodology.