Off-line identification of nonlinear systems using structurally adaptive radial basis function networks

Author

Junge, Thomas F. ; Unbehauen, Heinz

Author_Institution

Control Eng. Lab., Ruhr-Univ., Bochum, Germany

Volume

1

fYear

1996

fDate

11-13 Dec 1996

Firstpage

943

Abstract

This paper presents a novel off-line algorithm to train direct linear feedthrough radial basis function (DLF-RBF) networks. The algorithm basically explores the model error surfaces and combines an automatic determination of the number of RBF neurons with a hybrid optimization step to tune all parameters in the network. This leads to parsimonious models of SISO or MIMO dynamical systems, a primordial aim when solving nonlinear system identification problems. To demonstrate the effectiveness and the performance of the new method, it is applied to the identification of two highly nonlinear systems (one SISO and one MIMO system)

Keywords

MIMO systems; feedforward neural nets; identification; learning (artificial intelligence); nonlinear dynamical systems; MIMO dynamical systems; SISO dynamical systems; direct linear feedthrough radial basis function networks; hybrid optimization; model error surfaces; nonlinear systems; off-line identification; parsimonious models; structurally adaptive radial basis function networks; Adaptive systems; Artificial neural networks; MIMO; Multidimensional systems; Neurons; Nonlinear dynamical systems; Nonlinear systems; Programmable control; Radial basis function networks; System identification;

fLanguage

English

Publisher

ieee

Conference_Titel

Decision and Control, 1996., Proceedings of the 35th IEEE Conference on

Conference_Location

Kobe

ISSN

0191-2216

Print_ISBN

0-7803-3590-2

Type

conf

DOI

10.1109/CDC.1996.574586

Filename

574586