Skip to Main content Skip to Navigation

Surveillance et estimation de l'état de véhicules automobiles : un champ d'applications pour l'optimisation multi-objectif

Abstract : This thesis deals with multi-objective optimization for expensive evaluation functions and its application for the tuning of vehicle monitoring and driver assistance systems in the context of automotive engineering. After the presentation of a state of the art on multi-objective optimization methods, a new algorithm based on a double kriging strategy is proposed in order to reduce the number of objective function evaluations. The first application of the optimization algorithm concerns the tuning of an indirect tire-pressure monitoring system (without sensor). In addition to the tuning of high number of parameters, the optimization is constrained by the use of a large database, containing measurements obtained in real driving situations, and has to provide an optimal tuning in a limited processing time. The second application falls within lateral control of an autonomous driving system in a traffic jam situation. The current solution used by Renault/Nissan is based on a Kalman filter. In order to reduce the number of tuning parameters, solutions based on nonlinear approaches (high gain and sliding mode) are proposed. The tuning of the gains appearing in all these observers has been performed by using the proposed optimization algorithm, and by taking into account real driving data with measurements contained in a large database.
Complete list of metadata

Cited literature [173 references]  Display  Hide  Download
Contributor : Saïd Moussaoui Connect in order to contact the contributor
Submitted on : Wednesday, October 18, 2017 - 4:56:52 AM
Last modification on : Wednesday, April 27, 2022 - 3:50:06 AM
Long-term archiving on: : Friday, January 19, 2018 - 12:32:52 PM


Files produced by the author(s)


  • HAL Id : tel-01618471, version 1


Joan Davins-Valldaura. Surveillance et estimation de l'état de véhicules automobiles : un champ d'applications pour l'optimisation multi-objectif. Automatique / Robotique. Ecole Centrale de Nantes (ECN), 2016. Français. ⟨tel-01618471⟩



Record views


Files downloads