2020-05-01T16:07:05 n18178

EGAD - Evolved Grasping Analysis Dataset

Viewed: 8

The Evolved Grasping Analysis Dataset (EGAD), a collection of generated shapes for training and benchmarking robotic grasping and manipulatoin tasks. 

Diverse and extensive training data are critical to training modern robotic systems to grasp, and yet many systems are trained on small or non-diverse datasets repurposed from other domains. We used evoluationary algorithms to create a set of objects which uniformly span the object space of simple to complex, and easy to difficult to grasp, with a focus on geometric diversity. The objects are all easily 3D-printable, making 1:1 sim-to-real transfer possible.

Full details, code and videos can be found on the project website: https://dougsm.github.io/egad/

A preprint version of the paper can be found at: https://arxiv.org/abs/2003.01314

 

Geographical area of data collection

text
N/A

Publications

EGAD! an Evolved Grasping Analysis Dataset for diversity and reproducibility in robotic manipulation.
IEEE Robotics and Automation Letters (RA-L). Accepted April, 2020. Preprint Version: https://arxiv.org/abs/2003.01314

Research areas

Artificial Intelligence and Image Processing not elsewhere classified

Cite this collection

Morrison, Douglas; Corke, Peter; Leitner, Jurgen (2020): EGAD - Evolved Grasping Analysis Dataset for diversity and reproducibility in robotic manipulation. Queensland University of Technology. (Dataset) https://doi.org/10.25912/5eaa52c6eb6b4

Related information

Licence

BSD-3 (https://github.com/dougsm/egad/blob/master/LICENSE)

Copyright

© Doug Morrison, Australian Center for Robotic Vision, Queensland University of Technology, 2020

Connections

Contacts

Name: Doug Morrison

Other

Date record created:
2020-04-20T15:49:04
Date record modified:
2020-05-01T16:07:05
Record status:
Published - Open Access