Optimal length of data for identification of time varying system

Author

Tsumura, K. ; Oishi, Y.

Author_Institution

Dept. of Math. Eng. & Inf. Phys., Tokyo Univ., Japan

Volume

4

fYear

1999

fDate

6/21/1905 12:00:00 AM

Firstpage

3224

Abstract

In this paper, we analyze system uncertainties of time varying systems caused by the identification of the least squares method and show there exists an optimal length of data for identification. The errors between identification models and true systems at each time are investigated and we show that the radius of each corresponding uncertainty can be calculated in a statistical sense. We explain that these errors are composed of noise error term and a time varying term of systems. And it is shown that there exists a tradeoff between the sizes of these terms for the length of data in identification, and the optimal length of data which minimize the radius can be calculated

Keywords

identification; least squares approximations; minimisation; statistical analysis; time-varying systems; uncertain systems; corresponding uncertainty; least-squares method; noise error term; optimal data length; radius minimization; statistical calculation; system uncertainties; time varying system identification; Data engineering; Error analysis; Error correction; Information analysis; Least squares methods; Nonlinear systems; Physics; Robust control; Time varying systems; Uncertainty;

fLanguage

English

Publisher

ieee

Conference_Titel

Decision and Control, 1999. Proceedings of the 38th IEEE Conference on

Conference_Location

Phoenix, AZ

ISSN

0191-2216

Print_ISBN

0-7803-5250-5

Type

conf

DOI

10.1109/CDC.1999.827766

Filename

827766