On kernel design for regularized LTI system identification

2016/12/12 by Chen, Tianshi · 2 citations
#FOS: Electrical engineering #Systems and Control (eess.SY) #electronic engineering #information engineering

paper · doi:10.48550/arxiv.1612.03542

Abstract

There are two key issues for the kernel-based regularization method: one is how to design a suitable kernel to embed in the kernel the prior knowledge of the LTI system to be identified, and the other one is how to tune the kernel such that the resulting regularized impulse response estimator can achieve a good bias-variance tradeoff. In this paper, we focus on the issue of kernel design. Depending on the type of the prior knowledge, we propose two methods to design kernels: one is from a machine learning perspective and the other one is from a system theory perspective. We also provide analysis results for both methods, which not only enhances our understanding for the existing kernels but also directs the design of new kernels.

Cited by

• Numerically Efficient and Stable Algorithms for Kernel-Based Regularized System Identification Using Givens-Vector Representation
• Generalized collective quantum tomography: algorithm design, optimization, and validation

Related

• On Input Design for Regularized LTI System Identification: Power-constrained Input
• Kernels for linear time invariant system identification
• On Robustness of Kernel-Based Regularized System Identification
• Regularized linear system identification using atomic, nuclear and kernel-based norms: the role of the stability constraint
• Regularized system identification using orthonormal basis functions