Balanced identification and model reduction of a separable denominator 2-D system

Author

Treasure, R. ; Sreeram, V. ; Ngan, K.N.

Author_Institution

Sch. of Electr., Electron. & Comput. Eng., Western Australia Univ., Nedlands, WA, Australia

Volume

3

fYear

2004

fDate

20-23 July 2004

Firstpage

2048

Abstract

An identification and model reduction algorithm for linear discrete time invariant (LTDI) 2-D systems is proposed. A causal recursive separable denominator (CRSD) model structure is used. This allows for the 2-D system to be modeled as two independent 1-D systems. The system is then identified by extending the 1-D system identification Kung´s algorithm. Reduced order models are obtained by singular value decomposition (SVD) of the 2-D impulse response data. These are shown to converge to a balanced realization.

Keywords

discrete time systems; identification; linear systems; reduced order systems; singular value decomposition; Kung algorithm; balanced identification; causal recursive separable denominator model structure; impulse response; linear discrete time invariant 2D system; model reduction algorithm; separable denominator; singular value decomposition; Control system analysis; Control system synthesis; Controllability; Mathematical model; Observability; Reduced order systems; Singular value decomposition; Stability; State-space methods; System identification;

fLanguage

English

Publisher

ieee

Conference_Titel

Control Conference, 2004. 5th Asian

Conference_Location

Melbourne, Victoria, Australia

Print_ISBN

0-7803-8873-9

Type

conf

Filename

1426942