Regularized LTI System Identification with Multiple Regularization Matrix⁎

Tianshi Chen, Martin S. Andersen, Biqiang Mu, Feng Yin, Lennart Ljung, S. Joe Qin

Research output: Contribution to journalJournal articleResearchpeer-review

327 Downloads (Pure)

Abstract

Regularization methods with regularization matrix in quadratic form have received increasing attention. For those methods, the design and tuning of the regularization matrix are two key issues that are closely related. For systems with complicated dynamics, it would be preferable that the designed regularization matrix can bring the hyper-parameter estimation problem certain structure such that a locally optimal solution can be found efficiently. An example of this idea is to use the so-called multiple kernel Chen et al. (2014) for kernel-based regularization methods. In this paper, we propose to use the multiple regularization matrix for the filter-based regularization. Interestingly, the marginal likelihood maximization with the multiple regularization matrix is also a difference of convex programming problem, and a locally optimal solution could be found with sequential convex optimization techniques.
Original languageEnglish
Book seriesI F A C Workshop Series
Volume51
Issue number15
Pages (from-to)180-185
Number of pages6
ISSN1474-6670
DOIs
Publication statusPublished - 2018

Keywords

  • System identification
  • Regularization methods
  • Convex optimization

Fingerprint

Dive into the research topics of 'Regularized LTI System Identification with Multiple Regularization Matrix⁎'. Together they form a unique fingerprint.

Cite this