robotics
Human Dexterity Was Retargeted Without Copying the Hand
A generative retargeter maps human motion onto robot hands with different geometry and constraints.
Summary
A generative retargeter maps human motion onto robot hands with different geometry and constraints.
The method learns a distribution over robot motions compatible with a human demonstration instead of forcing one geometric correspondence. That creates room to respect joint limits and embodiment differences while preserving task intent. The evidence is limited to the authors’ tested hands and tasks; it does not make arbitrary human motion robot-safe.
Why it matters
A generative retargeter maps human motion onto robot hands with different geometry and constraints.
Limits and context
- The evidence is limited to the authors’ tested hands and tasks; it does not make arbitrary human motion robot-safe.
Key claims
A generative retargeter maps human motion onto robot hands with different geometry and constraints.
Qualification: The evidence is limited to the authors’ tested hands and tasks; it does not make arbitrary human motion robot-safe.
Evidence: source-2026-10-10-018
Sources
- arXiv preprint 2610.12440, original v1arXiv / submitting research authors · primary research
Corrections
No corrections have been recorded for this story.