Non-asymptotic state-space identification of closed-loop stochastic linear systems using instrumental variables

DOI10.1016/J.SYSCONLE.2023.105565zbMATH Open1520.93585arXiv2301.12537OpenAlexW4380201139MaRDI QIDQ6137801FDOQ6137801

Authors: Szabolcs Szentpéteri, Balázs Csanád Csáji 

Publication date: 4 September 2023

Published in: Systems \& Control Letters (Search for Journal in Brave)

Abstract: The paper suggests a generalization of the Sign-Perturbed Sums (SPS) finite sample system identification method for the identification of closed-loop observable stochastic linear systems in state space form. The solution builds on the theory of matrix-variate regression and instrumental variable methods to construct distribution-free confidence regions for the state space matrices. Both direct and indirect identification are studied, and the exactness as well as the strong consistency of the construction are proved. Furthermore, a new, computationally efficient ellipsoidal outer-approximation algorithm for the confidence regions is proposed. The new construction results in a semidefinite optimization problem which has an order-of-magnitude smaller number of constraints, as if one applied the ellipsoidal outer-approximation after vectorization. The effectiveness of the approach is also demonstrated empirically via a series of numerical experiments.

Full work available at URL: https://arxiv.org/abs/2301.12537

Recommendations

• Identification of non-linear stochastic systems by state dependent parameter estimation
• Instrumental variable methods for closed-loop system identification
• A General Framework for Nonparametric Identification of Nonlinear Stochastic Systems
• Closed-loop persistent identification of linear systems with unmodeled dynamics and stochastic disturbances
• Closed-loop system identification with recursive modifications of the instrumental variable method
• Identification of continuous-time linear dynamic stochastic systems
• Closed-loop system identification with modifications of the instrumental variable method
• scientific article; zbMATH DE number 227381
• Nonparametric identification of linear dynamic errors-in-variables systems
• On noncausal weighted least squares identification of nonstationary stochastic systems

zbMATH Keywords

instrumental variablesclosed-loop identificationdistribution-free methodsnon-asymptotic guarantees

Mathematics Subject Classification ID

Linear systems in control theory (93C05) Identification in stochastic control theory (93E12)

Cites Work

• CVXPY: a Python-embedded modeling language for convex optimization
• Title not available (Why is that?)
• Title not available (Why is that?)
• On covariance function tests used in system identification
• Title not available (Why is that?)
• Closed-loop identification revisited
• Shape Sensitivity Analysis and Gradient-Based Optimization of Large Structures Using MLFMA
• Guaranteed non-asymptotic confidence regions in system identification
• Finite sample properties of system identification methods
• Perturbed datasets methods for hypothesis testing and structure of corresponding confidence sets
• Non-Asymptotic Confidence Sets for the Parameters of Linear Transfer Functions
• Finite sample properties of linear model identification
• Asymptotic properties of SPS confidence regions
• Non-asymptotic confidence regions for model parameters in the presence of unmodelled dynamics
• Finite time identification in unstable linear systems
• Sign-Perturbed Sums: A New System Identification Approach for Constructing Exact Non-Asymptotic Confidence Regions in Linear Regression Models
• Facing undermodelling in sign-perturbed-sums system identification
• Guaranteed characterization of exact non-asymptotic confidence regions as defined by LSCR and SPS

Cited In (3)

• Highly-efficient filtered hierarchical identification algorithms for multiple-input multiple-output systems with colored noises
• Comment on 'Residual models and stochastic realization in state-space system identification' by R. Johansson et al.
• Instrumental variable scheme for closed-loop LPV model identification

This page was built for publication: Non-asymptotic state-space identification of closed-loop stochastic linear systems using instrumental variables

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6137801)