Kernel-based identification with frequency domain side-information

Khosravi, Mohammad; Smith, Roy S.

doi:10.1016/j.automatica.2022.110813

Cited by 4 publications

(2 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the standard kernels defined in (11), (12), and ( 13), one can obtain the closed-form of ϕ (u) τ using (17). To this end, we need ψ TC , ψ DC , and ψ SS , which are provided by the following theorem.…”

Section: B Optimization Problem Configuration: Continuous-time Casementioning

confidence: 99%

“…Various forms of side-information, such as stability, dissipativity, and region of attraction, are considered in identifying nonlinear dynamical systems [5], [6], [7], [8]. On the other hand, due to the importance of linear systems in practice, different sorts of side-information are included in their identification procedure, e.g., location of poles [9], structural properties [10], moments and derivatives of the transfer function [11], passivity [12], internal low complexity [13], positivity [14], stability [4], [15], [16], and generic frequency domain attributes [17].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Kernel-Based Impulse Response Identification With Side-Information on Steady-State Gain

Khosravi

Smith²

2023

IEEE Trans. Automat. Contr.

View full text Add to dashboard Cite

In this article, we consider the problem of system identification when side-information is available on the steadystate gain (SSG) of the system. We formulate a general nonparametric identification method as an infinite-dimensional constrained convex program over the reproducing kernel Hilbert space (RKHS) of stable impulse responses. The objective function of this optimization problem is the empirical loss regularized with the norm of RKHS, and the constraint is considered for enforcing the integration of the SSG side-information. The proposed formulation addresses both the discrete-time and continuous-time cases. We show that this program has a unique solution obtained by solving an equivalent finite-dimensional convex optimization. This solution has a closed-form when the empirical loss and regularization functions are quadratic and exact side-information is considered. We perform extensive numerical comparisons to verify the efficiency of the proposed identification methodology.

show abstract

Section: B Optimization Problem Configuration: Continuous-time Casementioning

confidence: 99%