Skip to Main content Skip to Navigation

Contribution à la navigation de robots mobiles : approche par modèle direct et commande prédictive

Abstract : Autonomous robots have to perform both control and perception tasks coordinately. Among these ones, the navigation task is a key in the interaction between the robot and its environment. It consists of determining the trajectories which the robot can follow in order to negotiate correctly around static and dynamic obstacles, assuming that it is programmed to map out its environment and situate itself within that environment. To perform this task, our approach rests on the direct kinematics model of the robot to generate admissible trajectories for the robot. Firstly, a reference trajectory is computed from the reference path provided by a path planer. Then the navigation task is modelized as an optimization under constraints problem, whose the cost function quantify the gap between the reference trajectory and the predicted trajectory of the robot. The obstacles are taken into account as constraints, and the minimization of the resulting cost function determinate the optimal control for the robot on a prediction horizon. This predictive navigation allows the robot to anticipate bi-pass movements on the chosen prediction horizon, Moreover, the behaviour of the robot is mastered by the use of parametered trajectories families.
Document type :
Complete list of metadatas

Cited literature [66 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Tuesday, September 14, 2010 - 2:21:08 PM
Last modification on : Thursday, March 5, 2020 - 6:49:01 PM
Long-term archiving on: : Wednesday, December 15, 2010 - 2:45:39 AM


Version validated by the jury (STAR)


  • HAL Id : tel-00517376, version 1


Nicolas Morette. Contribution à la navigation de robots mobiles : approche par modèle direct et commande prédictive. Autre. Université d'Orléans, 2009. Français. ⟨NNT : 2009ORLE2071⟩. ⟨tel-00517376⟩



Record views


Files downloads