Implementation and testing of point cloud based grasping algorithms for objetct picking
Ver/ Abrir
Metadatos
Mostrar el registro completo del ítemcomunitat-uji-handle:10234/158176
comunitat-uji-handle2:10234/71345
comunitat-uji-handle3:10234/174286
comunitat-uji-handle4:
TFG-TFMMetadatos
Título
Implementation and testing of point cloud based grasping algorithms for objetct pickingAutoría
Tutor/Supervisor; Universidad.Departamento
Morales Escrig, Antonio; Universitat Jaume I. Departament d'Enginyeria i Ciència dels ComputadorsFecha de publicación
2017-07-12Editor
Universitat Jaume IResumen
The purpose of this study is to investigate the most effective methodologies for the
grasping of items in an environment where success, robustness and time of the
algorithmic computation and its implementation are ... [+]
The purpose of this study is to investigate the most effective methodologies for the
grasping of items in an environment where success, robustness and time of the
algorithmic computation and its implementation are a key constraint. The study
originates from the Amazon Robotics Challenge 2017 (ARC’17) which addresses the
problem of automating the picking process in online shopping warehouses. In a real
warehouse environment the robot has to deal with restricted visibility and accessibility. The proposed solution to grasping was to retrieve a final position and orientation of the end effector given only sensory information without mesh reconstruction.
Two grippers were used: a two finger gripper with a narrow opening width and a
vacuum gripper. Antipodal Grasp Identification and Learning (AGILE) and Height
Accumulated Features (HAF) methods were chosen for implementation on a two
finger gripper due to their ease of applicability, same type of input, and reportedly
high success rate. One major contribution of this work was the creation of the Centroid Normals Approach (CNA) method for the vacuum gripper that chooses the
most central point cloud grasp location on the flattest part of the object. Since it does
not include calculation of orientation, its computation time is faster than the other
approaches. It was concluded that CNA should be used on as many objects as possible with both the vacuum gripper and the two finger gripper. A final scheme has
been devised to pick up the maximum number of items by combining algorithms
on the two different grippers, given the hardware restrictions, to cater to different
objects in the challenge. [-]
Palabras clave / Materias
Descripción
Treball de Final de Màster Universitari Erasmus Mundus en Robòtica Avançada. Codi: SJD024. Curs acadèmic 2016-2017
Tipo de documento
info:eu-repo/semantics/masterThesisDerechos de acceso
info:eu-repo/semantics/openAccess
Aparece en las colecciones
El ítem tiene asociados los siguientes ficheros de licencia: