Title :
Direct subspace model identification of LTI systems operating in closed-loop
Author :
Oku, Hiroshi ; Fujii, Takao
Author_Institution :
Dept. of Electron., Inf. & Commun. Eng., Osaka Inst. of Technol., Japan
Abstract :
In this paper, a subspace identification method of estimating the input-output relation of a linear system operating in closed-loop is presented. The proposed method is a straightforward consequence of the integration of the two-stage method with PI-MOESP type subspace identification method. It is shown that such integration brings the omission of the first stage identification, i.e., the omission of the identification of the sensitivity function of closed-loop and subsequent reconstruction of the fictitious input sequence. Therefore, the proposed method is able to estimate the input-output relation of a linear system in closed-loop directly from sampled sequences of observed signals.
Keywords :
closed loop systems; linear systems; parameter estimation; sensitivity analysis; signal processing; LTI systems; PI-MOESP type subspace identification method; closed-loop operation; direct subspace model identification; fictitious input sequence; input-output relation estimation; linear system; sensitivity function identification; signal sequences; subsequent sequence reconstruction; two-stage method; Colored noise; Control design; Error correction; Feedback control; Iterative methods; Linear systems; Power system modeling; Reduced order systems; Robust control; Signal processing;
Conference_Titel :
Decision and Control, 2004. CDC. 43rd IEEE Conference on
Conference_Location :
Nassau
Print_ISBN :
0-7803-8682-5
DOI :
10.1109/CDC.2004.1430378