Regularized LTI System Identification with Multiple Regularization Matrix⁎. Issue 15 (2018)

• Record Type: Journal Article
• Authors: Chen, Tianshi; Andersen, Martin S.; Mu, Biqiang; Yin, Feng; Ljung, Lennart; Qin, S. Joe
• Is Part Of: IFAC-PapersOnLine. Volume 51:Issue 15(2018)
• Page Start: 180
• Page End: 185
• Publication Date: 2018
• Subjects: System identification -- regularization methods -- sequential convex optimization; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2018.09.121
• Languages: English
• ISSNs: 2405-8963

Abstract

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.