Layer-wise dynamic event-triggered neural network control for discrete-time nonlinear systems - Université Toulouse 1 Capitole
Pré-Publication, Document De Travail Année : 2025

Layer-wise dynamic event-triggered neural network control for discrete-time nonlinear systems

Résumé

Recent efforts in the control community have focused on developing methods to guarantee closed-loop stability of systems controlled by multilayer perceptron (MLP)-based policies. However, little attention has been paid to the computational capacity demand of such controllers due to frequent network evaluations. These requirements can be prohibitive in practical control applications, particularly when implemented on microcontrollers with limited computational resources. In this paper, we address this challenge by proposing a dynamic Event-Triggering Mechanism (ETM) to reduce the computational burden. Specifically, we focus on the stabilization of discrete-time Lur'e systems with input saturation. The proposed strategy reduces the evaluation frequency of the layers of the neural controller while preserving stability guarantees. The ETM is constructed using Linear Matrix Inequality (LMI)-based conditions, which leverage the known properties of activation functions and employ Finsler's lemma to reduce conservativeness. Numerical results demonstrate the effectiveness of the proposed method, achieving significant computational savings compared to state-of-the-art solutions.
Fichier principal
Vignette du fichier
MLP_Trig.pdf (615.47 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04870932 , version 1 (07-01-2025)

Identifiants

  • HAL Id : hal-04870932 , version 1

Citer

Marco Sterlini, Samuele Zoboli, Sophie Tarbouriech. Layer-wise dynamic event-triggered neural network control for discrete-time nonlinear systems. 2025. ⟨hal-04870932⟩
0 Consultations
0 Téléchargements

Partager

More