FACTR 2: Learning External Force Sensing for Commodity Robot Arms Improves Policy Learning
2026-06-10 • Robotics
RoboticsArtificial IntelligenceMachine Learning
AI summaryⓘ
The authors created a way called NEXT to estimate the forces on robot arms without needing expensive force sensors. NEXT learns quickly using data from the robot moving freely and can guess forces almost as well as actual sensors. They also introduced FIRST, a training method that helps robots learn better by focusing more on moments before and during contact. Together, these tools let regular robot arms understand and use force information for tasks and teleoperation without extra hardware.
robot manipulationjoint torque estimationforce sensorsteleoperationbehavior cloningdata-driven methodsrobot learningforce feedbackrobot armpolicy training
Authors
Steven Oh, Jason Jingzhou Liu, Tony Tao, Philip Han, Kenneth Shaw, Satoshi Funabashi, Ruslan Salakhutdinov, Deepak Pathak
Abstract
Contact-rich manipulation requires force sensitivity, but many robot arms lack dedicated force sensors due to their high cost. We present Neural External Torque Estimation (NEXT), a data-driven method that estimates external joint torques without needing any dedicated force sensors. NEXT trains in 1 minute from only 10 minutes of free-motion data, yet achieves estimates comparable to dedicated joint-torque sensors. NEXT enables force-feedback teleoperation on low-cost arms and improves policy learning through Force-Informed Re-Sampling Training (FIRST), which up-samples pre-contact and contact segments during behavior cloning. Across five long-horizon tasks, FIRST outperforms prior force-aware policies by over 17% in task progress. Together, NEXT and FIRST bring force-aware teleoperation and policy learning to off-the-shelf robots without additional sensing hardware. Video results and code are available at https://jasonjzliu.com/factr2