Unified Hand Action Space Enables Cross-Embodiment Robotic Control
YuXiang_IRVL · x · 2026-07-11
A new method called UHAS (Unified Hand Action Space) has been proposed to solve the action representation challenge when training robot foundation models across different morphologies (such as robotic hands with varying mechanical structures or finger counts, and even humans).
Instead of learning latent actions, this approach relies on geometric principles: it maps the deformation of a canonical sphere to the control commands of different robots. Experiments show that a single policy can simultaneously control four completely different robotic hands to perform dexterous manipulation tasks.
Related event: UHAS Unifies Action Space for Cross-Embodiment Robotic Hands(3 posts)→
More from Embodied
- RTX Spark laptops impress analyst: 30B local model runs quiet with great battery life — Scobleizer · 2026-10-27
- Re-uploading a 239GB robot dataset moved only 2.6MB — inside a petabyte-scale data pipeline — mervenoyann · 2026-10-27
- ROSCon UK photo tour: BBC's whale robot, da Vinci surgical robot and the OG PR2 — kamathsblog · 2026-10-27
- Teardown of a humanoid robot Z-axis linear actuator, with Zmotor Q&A — MarwaEldiwiny · 2026-10-27
- GPT-6 Astra Beats Gemini 3.8 Flash in Paired Robot Test, Yet Neither Completes the Task — qinzytech · 2026-10-27
- Teardown of Zmotor LA500 humanoid linear actuator: 324 g, 500 N peak thrust — MarwaEldiwiny · 2026-10-27