Filtrage et commande basée sur un observateur pour les systèmes stochastiques

Abstract : This thesis deals with the filtering and control of nonlinear systems described by Itô stochastic differential equations whose diffusion is controlled by a noise which is multiplied with the state vector. In this manuscript, the goal is to relax the conditions of stability used in the literature using the almost sure exponential stability, also called exponential stability with probability equal to one. A new theorem on the almost sure exponential stability of the equilibrium point of a class of triangular nonlinear stochastic systems is proposed: the stability of the whole system is ensured by the stability of each decoupled subsystem. This theorem is applied to the filtering of stochastics systems with multiplicative noises. Conditions for asymptotic rejection of perturbations occurring in a stochastic differential equation with multiplicative noises have been proposed. The considered stability is the almost sure exponential one. A bound of the Lyapunov exponent ensures the almost sure convergence rate to zero for the state of the system. A bang-bang control law is synthesized for a class of stochastic nonlinear systems in two cases: (i) state feedback and (ii) measured output feedback with an observer. The used stability is the almost sure exponential one. The bounded real lemma is developed for stochastic algebro-differential systems with multiplicative noises and the Itô formula given for thèse systems. This approach has been used for the synthesis of an H-ihfinity measured output feedback control law with the exponential mean square stability. An observer for nonlinear stochastic algebro-differential systems was proposed using the almost sure exponential stability
Document type :
Theses
Complete list of metadatas

Cited literature [276 references]  Display  Hide  Download

https://hal.univ-lorraine.fr/tel-01751363
Contributor : Thèses Ul <>
Submitted on : Thursday, March 29, 2018 - 1:17:44 PM
Last modification on : Tuesday, April 10, 2018 - 1:27:43 AM

File

DDOC_T_2015_0013_BARBATA.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01751363, version 1

Collections

Citation

Asma Barbata. Filtrage et commande basée sur un observateur pour les systèmes stochastiques. Autre. Université de Lorraine, 2015. Français. ⟨NNT : 2015LORR0013⟩. ⟨tel-01751363⟩

Share

Metrics

Record views

57

Files downloads

110