The International Journal of Robotics Research· 2019Q1
Learning dexterous in-hand manipulation
- 1,607citations
- Q1SCImago
- 2019year
Short summary
Reinforcement learning policies trained entirely in simulation successfully transferred to a physical robot, enabling vision-based dexterous in-hand object reorientation without human demonstrations.
AI-generated from the title and abstract; the full text is not read.
Key points
- Reinforcement learning policies were trained entirely in a simulated environment.
- Policies successfully transferred to a physical Shadow Dexterous Hand for object reorientation.
- The method did not rely on human demonstrations, with complex manipulation behaviors emerging naturally.
- Randomization of physical properties like friction and object appearance in simulation aided transfer.
AI-generated from the title and abstract; the full text is not read.
Abstract
We use reinforcement learning (RL) to learn dexterous in-hand manipulation policies that can perform vision-based object reorientation on a physical Shadow Dexterous Hand. The training is performed in a simulated environment in which we randomize many of the physical properties of the system such as friction coefficients and an object’s appearance. Our policies transfer to the physical robot despite being trained entirely in simulation. Our method does not rely on any human demonstrations, but many behaviors found in human manipulation emerge naturally, including finger gaiting, multi-finger coordination, and the controlled use of gravity. Our results were obtained using the same distributed RL system that was used to train OpenAI Five. We also include a video of our results: https://youtu.be/jwSbzNHGflM .
The authors' abstract, as published at the source. The International Journal of Robotics Research, 2019 · DOI ↗
Continue with a free account
Ask the paper: 3 free questions a day about this paper; save it, get its citation, new summaries every day for your field. Takeaways are Premium.
Continue free on the webSign in with Google or Apple; no card needed. You come back to this paper.
On your phone:
Field: Control and Systems Engineering
Control and Systems EngineeringEngineering