A simple start-up procedure for canonical form state space identification, based on subspace approximation

Author

Ljung, Lennart

Author_Institution

Dept. of Electr. Eng., Linkoping Univ., Sweden

fYear

1991

fDate

11-13 Dec 1991

Firstpage

1333

Abstract

The authors describe a procedure for estimating parameters in multivariable linear systems, parameterized in pseudo-observability state-space canonical forms. The procedure is based on a standard four-stage instrumental variable method and subsequent operations in the Hankel matrix of estimated impulse responses. The main use of the procedure is for initialization of maximum likelihood/prediction error estimation methods

Keywords

approximation theory; linear systems; matrix algebra; maximum likelihood estimation; multivariable control systems; observability; parameter estimation; state estimation; state-space methods; Hankel matrix; canonical form state space identification; four-stage instrumental variable method; impulse responses; initialization; maximum likelihood/prediction error estimation methods; multivariable linear systems; parameter estimation; pseudo-observability; start-up procedure; state estimation; subspace approximation; Economic forecasting; Error analysis; Instruments; Linear systems; Maximum likelihood estimation; Observability; Parameter estimation; Predictive models; State estimation; State-space methods; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

Decision and Control, 1991., Proceedings of the 30th IEEE Conference on

Conference_Location

Brighton

Print_ISBN

0-7803-0450-0

Type

conf

DOI

10.1109/CDC.1991.261608

Filename

261608