IRUS Total

Transferring end-to-end visuomotor control from simulation to real world for a multi-stage task

File Description SizeFormat 
1707.02267.pdfAccepted version5.95 MBAdobe PDFView/Open
Title: Transferring end-to-end visuomotor control from simulation to real world for a multi-stage task
Authors: James, S
Davison, A
Johns, E
Item Type: Conference Paper
Abstract: End-to-end control for robot manipulation and grasping is emerging as an attractive alternative to traditional pipelined approaches. However, end-to- end methods tend to either be slow to train, exhibit little or no generalisability, or lack the ability to accomplish long-horizon or multi-stage tasks. In this paper, we show how two simple techniques can lead to end-to-end (image to velocity) execution of a multi-stage task, which is analogous to a simple tidying routine, without having seen a single real image. This involves locating, reaching for, and grasping a cube, then locating a basket and dropping the cube inside. To achieve this, robot trajectories are computed in a simulator, to collect a series of control velocities which accomplish the task. Then, a CNN is trained to map observed images to velocities, using domain randomisation to enable generalisation to real world images. Results show that we are able to successfully accomplish the task in the real world with the ability to generalise to novel environments, including those with dynamic lighting conditions, distractor objects, and moving objects, including the basket itself. We believe our approach to be simple, highly scalable, and capable of learning long-horizon tasks that have until now not been shown with the state-of-the-art in end-to-end robot control.
Issue Date: 13-Nov-2017
Date of Acceptance: 1-Sep-2017
URI: http://hdl.handle.net/10044/1/64944
Publisher: PMLR
Start Page: 334
End Page: 343
Journal / Book Title: Proceedings of Machine Learning Research
Volume: 78
Copyright Statement: © 2018 by the author(s). Available under a CC-BY Attribution Licence (http://creativecommons.org/licenses/by/4.0)
Conference Name: Conference on Robot Learning
Publication Status: Published
Start Date: 2017-11-13
Finish Date: 2017-11-15
Conference Place: Mountain View, California
Open Access location: https://arxiv.org/pdf/1707.02267.pdf
Appears in Collections:Computing