Grasping objects in cluttered environments
Nicolay, Pierre
Promoteur(s) : Boigelot, Bernard
Date de soutenance : 26-jui-2019/27-jui-2019 • URL permanente : http://hdl.handle.net/2268.2/6750
Détails
Titre : | Grasping objects in cluttered environments |
Titre traduit : | [fr] Saisir des objets dans un environnement encombré |
Auteur : | Nicolay, Pierre |
Date de soutenance : | 26-jui-2019/27-jui-2019 |
Promoteur(s) : | Boigelot, Bernard |
Membre(s) du jury : | Cornélusse, Bertrand
Detry, Renaud Wehenkel, Louis Geurts, Pierre Van Droogenbroeck, Marc |
Langue : | Anglais |
Nombre de pages : | 85 |
Mots-clés : | [en] grasping [en] machine learning [en] computer vision [en] robotic |
Discipline(s) : | Ingénierie, informatique & technologie > Sciences informatiques |
Organisme(s) subsidiant(s) : | Army Research Laboratory |
Centre(s) de recherche : | NASA/Jet Propulsion Laboratory, California Institute of Technology |
Intitulé du projet de recherche : | Robotics collaborative technology alliance |
Public cible : | Chercheurs Professionnels du domaine |
Institution(s) : | Université de Liège, Liège, Belgique |
Diplôme : | Master en ingénieur civil en informatique, à finalité spécialisée en "intelligent systems" |
Faculté : | Mémoires de la Faculté des Sciences appliquées |
Résumé
[en] One of the core challenges in todays robotic manipulation is the grasping problem. It consists in designing a mathematical model of the environment so that the robot is able to compute a hand and finger trajectories that yields a grasping configuration. In order for autonomous robots to interact with their environment, we need them to be able to grasp objects by understanding the environment surrounding them. We will thus present in this work a new solution for the grasping problem in cluttered environments. More specifically, we consider in this work the problem of detecting, planning and executing a grasp in a cluttered complex and uncontrolled environment. We consider here only a single RGB-D camera as input to our model. To solve this problem we use a hybrid model composed of a context segmentation and geometric model. The segmentation model, based on deep learning, captures information about where, and the geometric model, using a dictionary of grasping prototypes alongside a searching algorithm, tells us how to grasp. The segmentation network achieve a jiccard index value of 76.61\% on the training set and 56.93\% on the validation set. We achieve an overall grasp succes rate of 2 out of 3 grasps.
Fichier(s)
Document(s)
Description: -
Taille: 36.82 MB
Format: Adobe PDF
Description: -
Taille: 53.44 kB
Format: Adobe PDF
Annexe(s)
Description: -
Taille: 106.94 kB
Format: image/png
Description: -
Taille: 315.98 kB
Format: image/png
Description: -
Taille: 71.06 kB
Format: image/png
Description: -
Taille: 89.1 kB
Format: image/png
Description: -
Taille: 170.38 kB
Format: image/png
Citer ce mémoire
L'Université de Liège ne garantit pas la qualité scientifique de ces travaux d'étudiants ni l'exactitude de l'ensemble des informations qu'ils contiennent.