Physics guided learning improves torque sensing for robotic teleoperation
LETS: Low-Overhead External-Torque Sensing for Force-Feedback Bilateral Teleoperation
Robotics
Summary
Robotic arms used for remote control often struggle to accurately measure the forces they experience, which makes it hard to reflect those forces back to the human operator. The authors propose LETS, a method that uses the robot's physical model and a bit of learning to estimate these forces more accurately without extra expensive sensors. LETS trains quickly on a small amount of robot motion data and improves the quality of force feedback during teleoperation. This makes controlling the robot feel smoother and more precise for the user.
What this means in practice
- •For robotic system developers: Implement real-time, low-cost external torque estimation for improved haptic feedback in remote-controlled robots without needing extra force sensors.
- •For remote surgery technicians: Enable more stable and accurate force feedback during teleoperated medical procedures through better torque sensing on robotic instruments.
Authors
Zonghao Guo, Dongchen Guan, Yubo Wang, Peng Yang
Abstract
Bilateral teleoperation with force feedback is becoming increasingly important, and the quality of external-torque estimation is critical to this process. However, manufacturer-provided disturbance-torque estimates on commercial robotic arms are often inaccurate, while the use of end-effector force/torque sensors adds cost and payload. We propose LETS, a physics-guided learning approach that assigns rigid-body dynamics to recursive Newton-Euler inverse dynamics and reserves learning for only the robot-specific residual. LETS uses only the robot's URDF model and joint feedback, deliberately excluding control commands and thereby decoupling external-torque estimation from control-command inputs. It trains on 12 minutes of free-motion data in under one minute, without requiring a dedicated end-effector six-axis force/torque sensor. In bilateral teleoperation, LETS achieves real-time external-torque estimation that enables stable force feedback. Its free-space error is 69% lower than the manufacturer-provided estimate, substantially reducing coupled oscillations between leader and follower; its contact error is 28% lower, yielding more accurate haptic feedback.