Skip to Main content Skip to Navigation

Contrôle véhicule autonome. Contrôle robuste et haute performance pour permettre les manœuvres à haute dynamique des véhicules autonomes

Abstract : The work proposed in this thesis is in the context of autonomous driving. In particular, the objective is the development of a control law for path tracking of collision avoidance maneuvers for an autonomous vehicle.Several non-linear models of the vehicle, capable of representing its behavior in high dynamics maneuvers, are presented. The purpose is to obtain a model for the synthesis of the controllers. The different vehicle models proposed take into consideration the dynamics of the longitudinal, lateral and yaw vehicle speeds. That allows to use the models for the synthesis of controllers that deals simultaneously with vehicle longitudinal and lateral control. Moreover, a non-linear model for tire forces and the variable representation for load transfer have been used for the vehicle models. In fact, the representation of the non-linear behavior of the tires, influenced by the load transfer, is critical in high dynamics maneuvers. Some simulation results allow to compare the different vehicle models and to choose the model used for the controllers synthesis.A linear time-variant model is obtained through the linearization of the chosen non-linear model. The LPV polytopic and grid-based approaches are then used to define two LPV models.Several controllers, static and dynamic, have been developed using the two LPV models. These controllers combine the wheels steering ang torques to stabilize the vehicle and to guarantee the vehicle path tracking on a set of collision avoidance maneuvers. The synthesis of the controllers is done using robust and optimal control methods, through the resolution of optimization problems subjected to LMI constraints. The saturations of the control signals and of the tire forces are taken into consideration in the control synthesis in order to maximize the region of attraction of the system in closed loop.Several simulation results, obtained using a high representativity simulation model, allow to asses the closed loop system performances in presence of non-zero initial conditions and parameter dispersions.
Document type :
Complete list of metadata
Contributor : ABES STAR :  Contact
Submitted on : Thursday, July 21, 2022 - 4:13:11 PM
Last modification on : Tuesday, August 16, 2022 - 10:30:06 AM


Version validated by the jury (STAR)


  • HAL Id : tel-03735769, version 1


Dario Penco. Contrôle véhicule autonome. Contrôle robuste et haute performance pour permettre les manœuvres à haute dynamique des véhicules autonomes. Automatique / Robotique. Université Paris-Saclay, 2022. Français. ⟨NNT : 2022UPASG039⟩. ⟨tel-03735769⟩



Record views


Files downloads