Étude des concepts de filtrage robuste aux méconnaissances de modèles et aux pertes de mesures. Application aux systèmes de navigation

Abstract : To solve the problem of estimating the state of a system, it is necessary to have at one's disposal a model governing the dynamic of the state variables and to measure directly or indirectly all or a part of these variables. The work presented in this thesis deals with the estimation issue in the presence of model uncertainties and sensor losses. The first part of this work represents the synthesis of a state estimation device for nonlinear systems. It consists in selecting a state estimator and properly tuning it. Then, thanks to a criterion introduced for the occasion, it consists in algorithmically designing a hardware redundancy aiming at compensating for some sensor losses. The second part of this work deals with the conception of a sub-model compensating for some model uncertainties. This sub-model, designed by using the Allan variance, is usable by a Kalman filter. This work has been used to take into account some gyroscopical drifts in a GPS-INS integrated navigation based on a constrained Kalman filter. The results obtained, coming from experiments on two plane trajectories, showed a safe and robust behaviour of the proposed method
Complete list of metadatas

Cited literature [157 references]  Display  Hide  Download

https://hal.univ-lorraine.fr/tel-01748704
Contributor : Thèses Ul <>
Submitted on : Thursday, March 29, 2018 - 11:46:41 AM
Last modification on : Wednesday, April 11, 2018 - 1:27:55 AM
Long-term archiving on : Friday, September 14, 2018 - 7:11:28 AM

File

2008_SIRCOULOMB_V.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01748704, version 1

Collections

Citation

Vincent Sircoulomb. Étude des concepts de filtrage robuste aux méconnaissances de modèles et aux pertes de mesures. Application aux systèmes de navigation. Autre. Institut National Polytechnique de Lorraine, 2008. Français. ⟨NNT : 2008INPL093N⟩. ⟨tel-01748704⟩

Share

Metrics

Record views

22

Files downloads

34