Canonical Bayesian Linear System Identification

TL;DR

This paper introduces a canonical form approach for Bayesian linear system identification that resolves parameter non-identifiability, enabling efficient inference, meaningful priors, and robust uncertainty estimates, especially with limited data.

Contribution

It embeds canonical forms into Bayesian inference for LTI systems, fully capturing system dynamics and resolving identifiability issues, which was not achieved in prior methods.

Findings

Canonical forms improve computational efficiency.

Posteriors are more interpretable and well-behaved.

Method provides robust uncertainty estimates with limited data.

Abstract

Standard Bayesian approaches for linear time-invariant (LTI) system identification are hindered by parameter non-identifiability; the resulting complex, multi-modal posteriors make inference inefficient and impractical. We solve this problem by embedding canonical forms of LTI systems within the Bayesian framework. We rigorously establish that inference in these minimal parameterizations fully captures all invariant system dynamics (e.g., transfer functions, eigenvalues, predictive distributions of system outputs) while resolving identifiability. This approach unlocks the use of meaningful, structure-aware priors (e.g., enforcing stability via eigenvalues) and ensures conditions for a Bernstein--von Mises theorem -- a link between Bayesian and frequentist large-sample asymptotics that is broken in standard forms. Extensive simulations with modern MCMC methods highlight advantages over standard parameterizations: canonical forms achieve higher computational efficiency, generate interpretable and well-behaved posteriors, and provide robust uncertainty estimates, particularly from limited data.