Report
PolyUMI wins Best Device Track at CoRL 2026’s UMI Arena workshop, coauthor says
PolyUMI’s team describes a wireless gripper that records vision, touch and contact audio for robot learning.
TLDR
Coauthor Rickmer Krohn says PolyUMI won Best Device Track at CoRL 2026’s UMI Arena workshop. The project describes a sensing finger that moves from handheld gripper to robot hand in about ten minutes. In box-shaking tests, the team reports roughly 80% accuracy identifying gears and screws inside a closed box with contact audio, versus 44% using vision alone. The rattle is data.
Combined views
16
2 Sources, first seen ago

Comments