Covariance Stability Conditions for Linear Systems under Unbounded Additive and Parametric Uncertainties in Stochastic MPC applications
Résumé
This paper proposes new expressions of the error covariance dynamics for discrete-time linear systems affected by unbounded additive and parametric uncertainties. In tube-based stochastic MPC approaches, this dynamics is important for the computation of stochastic invariant sets that are used for time-varying constraints tightening. Furthermore, this paper addresses stability conditions of the error covariance dynamics
that can be used either in checking a posteriori the stability or to design the prestabilizing feedback gain depending on the level of uncertainties. Numerical examples are proposed to compare the evolution of the theoretical and the empirical error covariance.
Domaines
Automatique / RobotiqueOrigine | Fichiers produits par l'(les) auteur(s) |
---|