Hits:

Indexed by:期刊论文

Date of Publication:2015-03-01

Journal:SCIENCE CHINA-INFORMATION SCIENCES

Included Journals:SCIE、EI、Scopus

Volume:58

Issue:3

ISSN No.:1674-733X

Key Words:subspace identification method; correlation function estimates; closed-loop system; asymptotic properties; the system dynamics

Abstract:A novel subspace identification method based on correlation function which estimates a state-space system dynamics of unknown plant operating in closed-loop experimental condition is proposed in this paper. It is shown that the cross-correlation function of the output and external input signals are equal to the cross-correlation function of the input and external signals filtered through the system dynamics since noise signal has no correlation with the external input. The proposed algorithm is developed to obtain unbiased estimates of system matrices based on time-shifted invariance of the correlation function estimates. Later the algorithm is compared to other popular subspace methods in the simulation study and the results show the effectiveness of our method in the presence of colored noise and low signal-to-noise ratios.