An Automata Theoretic Approach to Observer Design for Switched Linear Systems - Archive ouverte HAL
Article Dans Une Revue Automatica Année : 2024

An Automata Theoretic Approach to Observer Design for Switched Linear Systems

Georges Aazan
Antoine Girard
Luca Greco
Paolo Mason

Résumé

We present an approach for designing asymptotic observers for discrete-time switched linear systems. We first give an automata theoretic characterization of switching signals containing an infinite number of reconstructible sequences, i.e. sequences allowing to estimate the state of the system. We show that such switching signals can be generated by a deterministic Büchi automaton whose construction is given in the paper. Then, we present a methodology to design switched observers. These observers have an internal discrete state variable whose dynamics is given by the transition map of the Büchi automaton. We then present two approaches to design observer gains such that the observer is convergent for all switching signals whose occurrence rate of reconstructible sequences is higher than a tunable threshold. The first approach gives an explicit construction of the observer gains while the second one is based on linear matrix inequalities. For switched systems with invertible state matrices, we show that the proposed observer structure is universal in the sense that it is always possible to design an observer of the proposed form. We use a simple example to illustrate our methodology and then consider a case study in which we design an observer for a multicellular converter.
Fichier principal
Vignette du fichier
final.pdf (588.54 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04503485 , version 1 (15-03-2024)

Identifiants

Citer

Georges Aazan, Antoine Girard, Luca Greco, Paolo Mason. An Automata Theoretic Approach to Observer Design for Switched Linear Systems. Automatica, 2024, 165, pp.111689. ⟨10.1016/j.automatica.2024.111689⟩. ⟨hal-04503485⟩
87 Consultations
97 Téléchargements

Altmetric

Partager

More