Stochastic Gain-Scheduled Control of Discrete-Time Systems Characterized by Random Polytopes

Yohei Hosoe (1) , Yuki Michitani (1) , Dimitri Peaucelle (2) , Tomomichi Hagiwara (1)

1 Kyoto University
2 LAAS-MAC - Équipe Méthodes et Algorithmes en Commande

This paper discusses stability analysis and gain-scheduled controller synthesis for discretetime non-stationary stochastic systems characterized by random polytopes. An associated deterministic time-varying parameter is assumed to be available online, and how to use its information in adjustment of controller gains is discussed in a framework of stochastic control using linear matrix inequalities. Conservativeness is reduced using our recent results and the S-variable approach.

Mots clés: Linear parameter-varying systems; stochastic systems; LMIs; random polytopes; Uncertainty; Time-varying systems; Symmetric matrices; Stochastic processes; State feedback; Robust control

Domaines: Automatique

Yohei Hosoe, Yuki Michitani, Dimitri Peaucelle, Tomomichi Hagiwara. Stochastic Gain-Scheduled Control of Discrete-Time Systems Characterized by Random Polytopes. IEEE Control Systems Letters, 2024, 8, pp.229 - 234. ⟨10.1109/LCSYS.2024.3366179⟩. ⟨hal-04448570⟩