Generation of synthetic AI training data for robotic grasp-candidate identification and evaluation in intralogistics bin-picking scenarios

Publikationstyp
Conference Paper
Date Issued
2022-06
Sprache
English
Author(s)
Holst, Dirk  orcid-logo
Schoepflin, Daniel  orcid-logo
Schüppstuhl, Thorsten  orcid-logo
Institut
Flugzeug-Produktionstechnik M-23  
TORE-DOI
10.15480/882.4904
TORE-URI
http://hdl.handle.net/11420/14498
Start Page
284
End Page
292
Citation
31st International Conference on Flexible Automation and Intelligent Manufacturing (FAIM 2022)
Contribution to Conference
31st International Conference on Flexible Automation and Intelligent Manufacturing, FAIM 2022  
Publisher DOI
10.1007/978-3-031-18326-3_28
Scopus ID
2-s2.0-85141843055
Publisher
Springer International Publishing AG
Robotic bin picking remains a main challenge for the wide enablement of industrial robotic tasks. While AI-enabled picking approaches are encouraging they repeatedly face the problem of data availability. The scope of this paper is to present a method that combines analytical grasp research with the field of synthetic data creation to generate individual training data for use-cases in intralogistics transportation scenarios. Special attention is given to systematic grasp finding for new objects and unknown geometries in transportation bins and to match the generated data to a real two-finger parallel gripper. The presented approach includes a grasping simulation in Pybullet to investigate the general tangibility of objects under uncertainty and combines these findings with a previously reported virtual scene generator in Blender, which generates AI-images of fully packed transport boxes, including depth maps and necessary annotations. This paper, therefore, contributes a synthesizing and cross-topic approach that combines different facets of bin-picking research such as geometric analysis, determination of tangibility of objects, grasping under uncertainty, finding grasps in dynamic and restricted bin-environments, and automation of synthetic data generation. The approach is utilized to generate synthetic grasp training data and to train a grasp-generating convolutional neural network (GG-CNN) and demonstrated on real-world objects.
Subjects
Automation
Bin picking
Grasp Simulation
Synthetic data
Tangibility
MLE@TUHH
DDC Class
600: Technik
Publication version
submittedVersion
Lizenz
https://creativecommons.org/licenses/by/4.0/
Loading...
Thumbnail Image
Name

978-3-031-18326-3_28.pdf

Size

1.01 MB

Format

Adobe PDF