From frequency response data to state-space models

Author

Gu, Guoxiang ; Chu, Cheng-Chih ; Khargonekar, Pramod P.

Author_Institution

Dept. of Electr. & Comput. Eng., Louisiana State Univ., Baton Rouge, LA, USA

Volume

2

fYear

1995

fDate

13-15 Dec 1995

Firstpage

1252

Abstract

This paper considers system identification in the frequency domain where the experimental data is a finite set of noisy frequency response measurement samples. The proposed algorithm consists of a combination of identification in ℋ_∞ and a modification of Ho/Kung´s algorithm to produce state-space models. The associated modeling error in ℋ_∞ norm is analyzed based on the a priori information on the unknown true system as well as the level of noise. Upper bounds are derived for the worst-case identification errors. The proposed algorithm is applied to a typical lightly damped system as studied in Gu and Khargonekar (1993) that yields impressive simulation results

Keywords

frequency response; identification; modelling; state-space methods; ℋ_∞ identification; Ho/Kung´s algorithm; frequency domain; frequency response data; lightly damped system; modeling error; noisy frequency response measurement samples; state-space models; system identification; unknown true system; upper bounds; worst-case identification errors; Contracts; Finite impulse response filter; Frequency domain analysis; Frequency measurement; Frequency response; Information analysis; Laboratories; Noise level; System identification; Upper bound;

fLanguage

English

Publisher

ieee

Conference_Titel

Decision and Control, 1995., Proceedings of the 34th IEEE Conference on

Conference_Location

New Orleans, LA

ISSN

0191-2216

Print_ISBN

0-7803-2685-7

Type

conf

DOI

10.1109/CDC.1995.480269

Filename

480269