Robot-Matched Soft Fingers for UMI Data Collection: Lessons from 100 Hours of Pick and Place
Overview
This research introduces a battery-powered handheld Universal Manipulation Interface (UMI) built on YUBI, featuring soft angled fingers that wrap around objects and deform visibly under load to serve as a visual proxy for contact forces. The interface matches the grippers of target robots, enabling efficient data collection for training multi-task pick-and-place policies. Using this system, the authors collected a 99.7-hour dataset across 53 objects, trained policies on 49 of them, and held out 4 as unseen test cases. Results indicate that adding a top camera improves success rates by 13 percentage points even when only half the data includes it, and mid-training on multi-task data boosts fine-tuning success on unseen objects from 35% to 57%. The finger design, dataset, and processing code are available at https://lute-corp.github.io/soft-yubi-site/.
Generated from attributed reports · Updated 2 hours ago
Event evidence and corrections
0 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Reported quantity · Reported duration: 99.7 hours · Basis not reported
“ery-powered handheld interface, built on YUBI, whose soft, angled fingers wrap around objects, deform visibly under load as a visual proxy for force, and match the grippers of the robot. With it, we collect a 99.7-hour dataset over 53 objects, train $\pi_{0.5}$ policies on 49 of them, and hold out the other 4 as unseen objects. We find that a top camera raises success by 13 points even when only about half of the dat”
Exact source · revision 1Source owner not reported
Artifact availability · code: reported pending
“a includes it, and that mid-training on our multi-task data raises success after fine-tuning on unseen objects from 35% to 57%. Finger design, dataset, and data processing code: https://lute-corp.github.io/soft-yubi-site/”
Exact source · revision 1Source owner not reported
Dataset · modalities: camera · Differing source assertions
“ery-powered handheld interface, built on YUBI, whose soft, angled fingers wrap around objects, deform visibly under load as a visual proxy for force, and match the grippers of the robot. With it, we collect a 99.7-hour dataset over 53 objects, train $\pi_{0.5}$ policies on 49 of them, and hold out the other 4 as unseen objects. We find that a top camera raises success by 13 points even when only about half of the dat”
Exact source · revision 1Source owner not reported
Dataset · modalities: force · Differing source assertions
“ery-powered handheld interface, built on YUBI, whose soft, angled fingers wrap around objects, deform visibly under load as a visual proxy for force, and match the grippers of the robot. With it, we collect a 99.7-hour dataset over 53 objects, train $\pi_{0.5}$ policies on 49 of them, and hold out the other 4 as unseen objects. We find that a top camera raises success by 13 points even when only about half of the dat”
Exact source · revision 1Source owner not reported
Dataset · tasks: pick and place
“arXiv:2611.03779v1 Announce Type: new Abstract: Training robot policies for multi-task pick and place requires demonstrations on many objects, which handheld Universal Manipulation Interfaces (UMIs) can collect quickly and cheaply. However, existing UMIs depend on external base stations, use rigid fingers that give no signal about contact forces, or use parallel-jaw fingers that limit grip strength. We present a batt”
Exact source · revision 1Source owner not reported
Report timeline
Follow attributed reports and material updates.
- arXiv Robotics — research abstractsSignalRobot-Matched Soft Fingers for UMI Data Collection: Lessons from 100 Hours of Pick and Place
This research introduces a battery-powered handheld Universal Manipulation Interface (UMI) built on YUBI, featuring soft angled fingers that wrap around objects and deform visibly under load to serve as a visual proxy for contact forces. The interface matches the grippers of target robots, enabling efficient data collection for training multi-task pick-and-place policies. Using this system, the authors collected a 99.7-hour dataset across 53 objects, trained policies on 49 of them, and held out 4 as unseen test cases. Results indicate that adding a top camera improves success rates by 13 percentage points even when only half the data includes it, and mid-training on multi-task data boosts fine-tuning success on unseen objects from 35% to 57%. The finger design, dataset, and processing code are available at https://lute-corp.github.io/soft-yubi-site/.
Event coverage history
There is not enough continuous observation data to show a trend.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).