One Contact Skill Transfers Dexterous Manipulation Across Robot Hands

GOTT separates reaching, contact acquisition and object motion tracking, allowing one feedback-driven contact policy to support varied tasks, objects and robot embodiments.

Big Tech
Yulin Liu · Lai Wei · Yen-Jen Wang · Akash Sharma · Pieter Abbeel · Henrik I. Christensen · +1 more

Amazon FAR · UC San Diego · University of Chicago · UC Berkeley

Research Digest··2 min read
Liu and colleagues introduce GOTT, a dexterous manipulation framework that executes robot-agnostic object trajectories using a reusable contact-acquisition skill.

GOTT represents manipulation intent with two inputs: a desired object trajectory and a reach specification that guides the hand toward a task-relevant contact region.

Why this paper

From Amazon FAR and 3 others

In one line

A reusable cross-embodiment contact-acquisition primitive enables dexterous manipulation from object trajectories and reach specifications.

What we could check

  • ·No code link found
  • ·No weights link found
  • ·No dataset link found
  • ·No compute details found
  • ✓Limitations stated by the authors (3 noted)
  • ·No benchmark numbers found

Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.

§

Research Digest

Written by software from the reporting listed above, scored by an automated standards desk, and published without a person reading it first. If something here is wrong, tell the editor and it will be put right.