DocumentCode :
3475625
Title :
Subspace algorithms for the stochastic identification problem
Author :
Van Overschee, Peter ; De Moor, Bart
Author_Institution :
Dept. of Electr. Eng., Katholieke Univ. Leuven, Belgium
fYear :
1991
fDate :
11-13 Dec 1991
Firstpage :
1321
Abstract :
The authors derive a novel algorithm to consistently identify stochastic state space models from given output data without forming the covariance matrix and using only semi-infinite block Hankel matrices. The algorithm is based on the concept of principle angles and directions. The authors describe how these can be calculated with only QR and QSVD decompositions. They also provide an interpretation of the principle directions as states of a non-steady-state Kalman filter. With a couple of examples, it is shown that the proposed algorithm is superior to the classical canonical correlation algorithms
Keywords :
Kalman filters; identification; matrix algebra; state-space methods; stochastic processes; QR decompositions; QSVD decompositions; covariance matrix; nonsteady-state Kalman filter; semi-infinite block Hankel matrices; stochastic identification; stochastic state space models; Covariance matrix; Gaussian noise; H infinity control; Least squares approximation; Matrix decomposition; Robustness; State estimation; State-space methods; Stochastic processes; Stochastic resonance;
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.1991.261604
Filename :
261604
Link To Document :
بازگشت