FOMI: A Force-Feedback Interface for Teaching Contact-Rich Manipulation
Overview
FOMI enables scalable human demonstration collection for policy learning using force feedback, improving efficiency by 29% and 18% in contact-rich tasks through robot-free and teleoperation support.
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.
Control: teleoperated
“arXiv:2603.06280v2 Announce Type: replace Abstract: Long-horizon mobile manipulation needs whole-body demonstrations, and collecting them is the bottleneck. Keeping the robot in the loop ties the dataset to robot uptime, while recording human motion without it instead requires a conversion step that decides the joint-space action the policy learns. SuperSuit is a single wearable that both teleoperates a wheeled biman”
Exact source · revision 1Source owner not reported
Reported quantity · robot tasks: 3 other · Basis not reported
“pers, with arm targets expressed as forward-looking increments of the observed joint state. On three real-robot tasks this lets active demonstrations replace most robot-in-the-loop collection without changing policy quality. With a 10-episode robot-grounded seed held fixed, the remaining 100 episodes can be recorded without the robot, whereas active-only training falls well short. SuperSuit moves the limit on dataset”
Exact source · revision 1Source owner not reported
Reported quantity · episodes: 100 other · Basis not reported · Differing source assertions
“pers, with arm targets expressed as forward-looking increments of the observed joint state. On three real-robot tasks this lets active demonstrations replace most robot-in-the-loop collection without changing policy quality. With a 10-episode robot-grounded seed held fixed, the remaining 100 episodes can be recorded without the robot, whereas active-only training falls well short. SuperSuit moves the limit on dataset”
Exact source · revision 1Source owner not reported
Reported quantity · episodes: 10 other · Basis not reported · Differing source assertions
“pers, with arm targets expressed as forward-looking increments of the observed joint state. On three real-robot tasks this lets active demonstrations replace most robot-in-the-loop collection without changing policy quality. With a 10-episode robot-grounded seed held fixed, the remaining 100 episodes can be recorded without the robot, whereas active-only training falls well short. SuperSuit moves the limit on dataset”
Exact source · revision 1Source owner not reported
Reported quantity · contact-rich tasks: 4 other · Basis not reported
“ction. By further mapping pose changes and gripper opening of FOMI to robot end-effector poses and gripper commands, the system also supports teleoperation. Across four contact-rich tasks, FOMI improves collection efficiency by approximately 29% and 18% over the best baselines in robot-free collection and teleoperation, respectively. Demonstrations collected with FOMI in both modes enable policies to perform challeng”
Exact source · revision 1Source owner not reported
Reported quantity · robot-free collection tasks: 4 other · Basis not reported
“ction. By further mapping pose changes and gripper opening of FOMI to robot end-effector poses and gripper commands, the system also supports teleoperation. Across four contact-rich tasks, FOMI improves collection efficiency by approximately 29% and 18% over the best baselines in robot-free collection and teleoperation, respectively. Demonstrations collected with FOMI in both modes enable policies to perform challeng”
Exact source · revision 1Source owner not reported
Report timeline
Follow attributed reports and material updates.
- arXiv Robotics — research abstractsFOMI: A Force-Feedback Manipulation Interface for Teaching Contact-Rich Manipulation
FOM, a force-feedback manipulation interface, enables scalable acquisition of human demonstrations for policy learning through intuitive force feedback. The system supports both robot-free collection and teleoperation, improving collection efficiency by approximately 29% and 18% over existing baselines in contact-rich tasks.
- arXiv Robotics — research abstractsSuperSuit: An Isomorphic Bimodal Interface for Scalable Mobile Manipulation
SuperSuit is a wearable device that teleoperates a wheeled bimanual robot and records demonstrations when the robot is absent. By mechanically aligning the wearable's arms with the robot's, it eliminates the need for inverse kinematics and retargeting. The system generates 22-D commands for the torso, planar base, arms, and grippers, enabling active demonstrations to replace most robot-in-the-loop data collection without affecting policy quality.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).