Title :
System identification using LQG-balanced model reduction
Author_Institution :
Dept. Autom. Control, Lund Univ., Sweden
Abstract :
System identification of linear multivariable dynamic models based on discrete-time data can be performed using a algorithm combining linear regression and LQG-balanced model reduction. The approach is applicable also to unstable system dynamics and it provides balanced models for optimal linear prediction and control.
Keywords :
discrete time systems; identification; linear quadratic Gaussian control; linear systems; multivariable systems; prediction theory; reduced order systems; stability; statistical analysis; LQG-balanced model reduction; discrete-time data; linear multivariable dynamic models; linear regression; optimal linear control; optimal linear prediction; system identification; unstable system dynamics; Automatic control; Colored noise; Linear systems; Maximum likelihood estimation; Optimal control; Optimization methods; Predictive models; Reduced order systems; Riccati equations; System identification;
Conference_Titel :
Decision and Control, 2002, Proceedings of the 41st IEEE Conference on
Print_ISBN :
0-7803-7516-5
DOI :
10.1109/CDC.2002.1184501
