Novel parameter estimation method for the systems with colored noises by using the filtering identification idea

Compared with the systems with white noise disturbances, the parameter identification of the systems with colored noises (i.e., correlated noises) is more difficult. In this letter, we use the model transformation to study the identification problem for the systems with colored noises by using the f...

Full description

Saved in:
Bibliographic Details
Published inSystems & control letters Vol. 186; p. 105774
Main Authors Xu, Ling, Ding, Feng, Zhang, Xiao, Zhu, Quanmin
Format Journal Article
LanguageEnglish
Published Elsevier B.V 01.04.2024
Subjects
Auxiliary model
Parameter estimation
Recursive identification
Stochastic gradient
Recursive identification
Parameter estimation
Auxiliary model
Stochastic gradient
Online AccessGet full text

Cover

More Information
Summary:Compared with the systems with white noise disturbances, the parameter identification of the systems with colored noises (i.e., correlated noises) is more difficult. In this letter, we use the model transformation to study the identification problem for the systems with colored noises by using the filtering identification idea. The basic idea is to transform a system with colored noise into two identification models with white noises and then to propose a novel parameter estimation algorithms, which can generate more accurate parameter estimates than some related existing identification algorithm. The proposed method can be applied to other linear or nonlinear stochastic systems with different structures and colored disturbance noises. The provided simulation example tests the proposed algorithm.
ISSN:0167-6911
1872-7956
DOI:10.1016/j.sysconle.2024.105774