Article: Support kernels regression for NARMA system identification (SKRNARMA) Journal: International Journal of Modelling, Identification and Control (IJMIC) 2014 Vol.22 No.2 pp.139 - 149 Abstract: This paper deals with the identification of nonlinear systems using multi-kernel approach. First, we have improved the support vector regression (SVR) method in order to identify nonlinear complex system. Our idea consists of dividing the regressor vector in several blocks, and, for each one a kernel function is used. This blockwise SVR approach is called support kernel regression (SKR). Furthermore, we have proposed two methods, SKR(rbf-lin) and SKR(rbf-rbf). Second, the SKR approach is improved to deal with the problem of NARMA system identification. Therefore, a new method called support kernel regression for NARMA model (SKRNARMA) is suggested. The basic idea is to consider the terms of auto-correlation and cross-correlation of the nonlinearity of input output discrete time processes, and for every term a kernel function is used. An example of MIMO system is presented for qualitative comparison with the classical SVR approach based on a single kernel function. The results reveal the accuracy and the robustness of the obtained model based on our proposed (SKRNARMA)-based approach. Inderscience Publishers - linking academia, business and industry through research

Title: Support kernels regression for NARMA system identification (SKR_NARMA)

Authors: Mounira Tarhouni; Salah Zidi; Kaouther Laabidi; M. Ksouri-Lahmari

Addresses: Laboratory of Research Analysis, Conception and Control of Systems (ACCS, ENIT), BP 37, Le Belvedere 1002, Tunis, Tunisia ' LAGIS (USTL, Lille), Villeneuve d'Ascq 59650 Lille, France ' Laboratory of Research Analysis, Conception and Control of Systems (ACCS, ENIT), BP 37, Le Belvedere 1002, Tunis, Tunisia ' Laboratory of Research Analysis, Conception and Control of Systems (ACCS, ENIT), BP 37, Le Belvedere 1002, Tunis, Tunisia

Abstract: This paper deals with the identification of nonlinear systems using multi-kernel approach. First, we have improved the support vector regression (SVR) method in order to identify nonlinear complex system. Our idea consists of dividing the regressor vector in several blocks, and, for each one a kernel function is used. This blockwise SVR approach is called support kernel regression (SKR). Furthermore, we have proposed two methods, SKR(rbf-lin) and SKR(rbf-rbf). Second, the SKR approach is improved to deal with the problem of NARMA system identification. Therefore, a new method called support kernel regression for NARMA model (SKR_NARMA) is suggested. The basic idea is to consider the terms of auto-correlation and cross-correlation of the nonlinearity of input output discrete time processes, and for every term a kernel function is used. An example of MIMO system is presented for qualitative comparison with the classical SVR approach based on a single kernel function. The results reveal the accuracy and the robustness of the obtained model based on our proposed (SKR_NARMA)-based approach.

Keywords: support vector regression; SVR; support kernel regression; SKR; NARMA model; nonlinear auto regressive moving average; nonlinear systems; system identification; kernel function; support vector machines; SVM.

DOI: 10.1504/IJMIC.2014.064289

International Journal of Modelling, Identification and Control, 2014 Vol.22 No.2, pp.139 - 149

Published online: 27 Sep 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Support kernels regression for NARMA system identification (SKRNARMA)

Keep up-to-date

Title: Support kernels regression for NARMA system identification (SKR_NARMA)