Nonlinear modelling and identification

There has been a considerable increase in activity in the field of identification of nonlinear systems. Side by side with the identification of discrete-time models based on Kolmogorov-Gabor polynomials, artificial neural networks, etc., there has been a great deal of progress in the identification of continuous-time models governed by ordinary differential equations. This paper attempts to give an overview of the existing modelling frameworks and makes a comparison among them on the basis of their approximating abilities, computational requirements, on-line applicability etc