Abstract:The focus of this paper is on linear system identification in the setting where it is known that the underlying partially-observed linear dynamical system lies within a finite collection of known candidate models. We first consider the problem of identification from a given trajectory, which in this setting reduces to identifying the index of the true model with high probability. We characterize the finite-time sample complexity of this problem by leveraging recent advances in the non-asymptotic analysis of linear least-square methods in the literature. In comparison to the earlier results that assume no prior knowledge of the system, our approach takes advantage of the smaller hypothesis class and leads to the design of a learner with a dimension-free sample complexity bound. Next, we consider the switching control of linear systems, where there is a candidate controller for each of the candidate models and data is collected through interaction of the system with a collection of potentially destabilizing controllers. We develop a dimension-dependent criterion that can detect those destabilizing controllers in finite time. By leveraging these results, we propose a data-driven switching strategy that identifies the unknown parameters of the underlying system. We then provide a non-asymptotic analysis of its performance and discuss its implications on the classical method of estimator-based supervisory control.

A Class of Linear Convergent Algorithms for H∞ Identification

H-Infinity Control for Linear Systems with State Saturation Nonlinearities

New time-domain algorithms for H∞ identification

An Algebra Approach in H_∞ System Identification

System identification: New developments of the asymptotic theory/method

Identification in Hh∞ Via Convex Programming

Identification in hH<sup>∞</sup> via convex programming

New closed-loop identification approach based on output over-sampling scheme

On Anti-Aliasing Filtering and Over-Sampling Scheme in System Identification

A LMI Approach to Robust H/sub ∞/ Filtering for Discrete-Time Systems

Simple Control-Relevant Identification Test Methods for A Class of Ill-Conditioned Processes

Identification of Hammerstein Models for Control

Comparison of Three Classes of Identification Methods

A Linearly Convergent Iterative Method for Identifying H-matrices

A Class of Convex Optimization-Based Recursive Algorithms for Identification of Stochastic Systems

One Step Iterative Strategy for Nonlinear System Identification

Optimal H∞ output feedback control for a class of nonlinear systems

Closed-loop Subspace Identification for Continuous-time Models

A general central estimation algorithm and the special cases in l 1 robust identification

A least-square method for non-asymptotic identification in linear switching control

H ∞ Reference Tracking Control Design for a Class of Nonlinear Systems with Time-Varying Delays