Sample Complexity Bounds for Linear System Identification From a Finite Set

被引：0

作者：

Chatzikiriakos, Nicolas ^{[1
]}

Iannelli, Andrea ^{[1
]}

机构：

[1] Institute for Systems Theory and Automatic Control, University of Stuttgart, Stuttgart

来源：

IEEE Control Systems Letters | 2024年 / 8卷

关键词：

finite sample analysis; Linear system identification; maximum likelihood estimation;

D O I：

10.1109/LCSYS.2024.3514995

中图分类号：

学科分类号：

摘要：

This letter considers a finite sample perspective on the problem of identifying an LTI system from a finite set of possible systems using trajectory data. To this end, we use the maximum likelihood estimator to identify the true system and provide an upper bound for its sample complexity. Crucially, the derived bound does not rely on a potentially restrictive stability assumption. Additionally, we leverage tools from information theory to provide a lower bound to the sample complexity that holds independently of the used estimator. The derived sample complexity bounds are analyzed analytically and numerically. © 2024 IEEE.

引用

页码：2751 / 2756

页数：5