Recursive Weighted Null-Space Fitting Method for Identification of Multivariate Systems

Fang, MengyuanZhejiang Sci Tech Univ, Zhejiang Engn Lab Ind Internet Commun Technol, Hangzhou, Peoples R China.;Northeastern Univ, State Key Lab Synthet Automat Proc Ind, Shenyang, Peoples R China. (author)

QC 20211022
Recursive identification of structured multivariate models is known to be difficult due to the general non-convexity of the likelihood function. In this work, we propose a recursive multivariate weighted null-space fitting method for identification of structured multivariate models. The proposed method first uses recursive least squares to estimate a high order non-parametric model, then a parametric model is obtained through weighted least squares from the non-parametric model. In this way, the method avoids directly optimizing a non-convex likelihood function and has guaranteed global convergency. Moreover, the proposed method is flexible in model structures and has the same finite sample performance as its off-line counterpart. We use simulation examples to illustrate the performance.

Galrinho, MiguelOHB Sweden AB, Kista, Sweden. (author)
Hjalmarsson, Håkan,1962-KTH,Reglerteknik(Swepub:kth)u10a8l40 (author)
Zhejiang Sci Tech Univ, Zhejiang Engn Lab Ind Internet Commun Technol, Hangzhou, Peoples R China.;Northeastern Univ, State Key Lab Synthet Automat Proc Ind, Shenyang, Peoples R China.OHB Sweden AB, Kista, Sweden. (creator_code:org_t)

By the author/editor: Fang, Mengyuan; Galrinho, Miguel; Hjalmarsson, Håk ...

About the subject

ENGINEERING AND TECHNOLOGY: ENGINEERING AND ...; and Electrical Engin ...; and Control Engineer ...

Kungliga biblioteket hanterar dina personuppgifter i enlighet med EU:s dataskyddsförordning (2018), GDPR. Läs mer om hur det funkar här.
Så här hanterar KB dina uppgifter vid användning av denna tjänst.