© The Institution of Electrical Engineers
A 2-step least-squares technique is extended to deal with the problem of estimating the parameters of the transfer-function matrix of a multivariable linear discrete system. The method does not require assumptions such as common dynamics and a single noise source. Results on its performance are presented.
References
-
-
1)
-
D.Q. Mayne
.
(1972)
Parametrization and identification of linear multivariable systems, Lecture notes in mathematics—Vol. 294.
-
2)
-
R.L. Kashyap ,
R.E. Nasburg
.
Parameter estimation in multivariate stochastic difference equations.
IEEE Trans.
,
784 -
797
-
3)
-
Abaza, B.A., Denham, M.J.: `A two-step least squares parameter estimation method: test case examples', Presented at the Round Table Discussion, IAFC Congress, 1975, Boston, USA.
-
4)
-
Pandya, R.N., Pagurek, B.: `Two stage least squares estimators and their recursive approximations', Proceedings of 3rd IFAC symposium on identification and system parameter estimation, 1973, The Hague Delft, The Netherlands, p. 701–710.
-
5)
-
M.J. Denham
.
Canonical forms for the identification of multivariable linear systems.
IEEE Trans.
,
646 -
656
-
6)
-
B.W. Dickinson ,
T. Kailath ,
M. Morf
.
Canonical matrix fraction and state-space descriptions for deterministic and stochastic linear systems.
IEEE Trans.
,
656 -
667
-
7)
-
M.J. Denham ,
B.A. Abaza
.
, System identification package.
http://iet.metastore.ingenta.com/content/journals/10.1049/el_19760254
Related content
content/journals/10.1049/el_19760254
pub_keyword,iet_inspecKeyword,pub_concept
6
6