Set-membership identification of continuous-time MIMO systems via Tustin discretization

TL;DR

This paper introduces a set-membership identification method for continuous-time MIMO systems using Tustin discretization, effectively handling bounded errors and derivative estimation issues.

Contribution

It presents a novel approach combining set-membership techniques with Tustin discretization to improve continuous-time system identification from sampled data.

Findings

Method reduces derivative estimation errors.

Approach formulates as polynomial optimization problem.

Numerical results validate effectiveness on simulated and experimental data.

Abstract

In this paper, we deal with the identification of continuous-time systems from sampled data corrupted by unknown but bounded errors. A significant challenge in continuous-time identification is the estimation of the input and output data derivatives. In this paper, we propose a novel method based on set-membership techniques and Tustin discretization, which overcomes the derivative measurement problem and the presence of bounded errors affecting all the measured signals. First, we derive the proposed method and prove that it becomes an affordable polynomial optimization problem. Then, we present some numerical results based on simulation and experimental data to explore the effectiveness of the proposed method.