VR-Goggles for Robots : Real-to-Sim Domain Adaptation for Visual Control

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review

27 Scopus Citations
View graph of relations

Author(s)

  • Jingwei Zhang
  • Lei Tai
  • Peng Yun
  • Yufeng Xiong
  • Ming Liu
  • And 2 others
  • Joschka Boedecker
  • Wolfram Burgard

Detail(s)

Original languageEnglish
Article number8620258
Pages (from-to)1148-1155
Journal / PublicationIEEE Robotics and Automation Letters
Volume4
Issue number2
Online published21 Jan 2019
Publication statusPublished - Apr 2019
Externally publishedYes

Abstract

In this letter, we deal with the reality gap from a novel perspective, targeting transferring deep reinforcement learning (DRL) policies learned in simulated environments to the real-world domain for visual control tasks. Instead of adopting the common solutions to the problem by increasing the visual fidelity of synthetic images output from simulators during the training phase, we seek to tackle the problem by translating the real-world image streams back to the synthetic domain during the deployment phase, to make the robot feel at home. We propose this as a lightweight, flexible, and efficient solution for visual control, as first, no extra transfer steps are required during the expensive training of DRL agents in simulation; second, the trained DRL agents will not be constrained to being deployable in only one specific real-world environment; and third, the policy training and the transfer operations are decoupled, and can be conducted in parallel. Besides this, we propose a simple yet effective shift loss that is agnostic to the downstream task, to constrain the consistency between subsequent frames which is important for consistent policy outputs. We validate the shift loss for artistic style transfer for videos and domain adaptation, and validate our visual control approach in indoor and outdoor robotics experiments.

Research Area(s)

  • Deep learning in robotics and automation, model learning for control, visual-based navigation

Citation Format(s)

VR-Goggles for Robots : Real-to-Sim Domain Adaptation for Visual Control. / Zhang, Jingwei; Tai, Lei; Yun, Peng; Xiong, Yufeng; Liu, Ming; Boedecker, Joschka; Burgard, Wolfram.

In: IEEE Robotics and Automation Letters, Vol. 4, No. 2, 8620258, 04.2019, p. 1148-1155.

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review