Static Output Feedback Synthesis of Time-Delay Linear Systems via Deep Unfolding

TL;DR

This paper introduces a novel deep unfolding approach combining neural networks and linear matrix inequalities to design static output feedback gains for stabilizing time-delay linear systems, enhancing control design methods.

Contribution

It presents a new algorithm that integrates neural ODEs with control theory for static output feedback synthesis in time-delay systems.

Findings

The proposed method effectively stabilizes time-delay systems in simulations.

Deep unfolding improves the design process by combining learning and control verification.

Numerical results demonstrate the algorithm's potential in control system stabilization.

Abstract

We propose a deep unfolding-based approach for stabilization of time-delay linear systems. Deep unfolding is an emerging framework for design and improvement of iterative algorithms and attracting significant attentions in signal processing. In this paper, we propose an algorithm to design a static output feedback gain for stabilizing time-delay linear systems via deep unfolding. Within the algorithm, the learning part is driven by NeuralODE developed in the community of machine learning, while the gain verification is performed with linear matrix inequalities developed in the systems and control theory. The effectiveness of the proposed algorithm is illustrated with numerical simulations.