Tactile-Proprioceptive Sensor Fusion for Contact Wrench Estimation in Whole-Body Physical Human-Robot Interaction
Junha Min, Junghyeon Ma, Jiwung Kwon, Sunggyu Bae, Joohyung Kim, Kyungseo Park
AI summary
Problem
Motor-current-based force estimation suffers from large friction-induced dead bands and hysteresis during stationary and quasi-static phases, degrading responsiveness in physical human-robot interaction. Traditional solutions rely on costly external sensors or ad-hoc thresholding that sacrifices sensitivity.
Approach
The system uses tactile skin signals to detect contact onset and cancel static friction, then applies a temporal convolutional network to model and compensate for friction hysteresis during motion transitions, fusing these with motor-current data for multi-axis force reconstruction.
Key results
- Tactile cues disambiguate static friction from external forces, eliminating estimation dead bands
- TCN compensator reduces static-to-kinetic friction residual RMSE by 54.8%
- Multi-axis contact forces are accurately reconstructed on the robot surface
- Demonstrated smooth, responsive kinesthetic teaching with improved sensitivity over tactile-only and proprioceptive-only baselines
Why it matters
Provides a scalable, low-cost pathway to safe and intuitive physical human-robot interaction by replacing expensive force sensors with fused tactile and proprioceptive data.
Abstract
Direct physical guidance is a natural means of teaching and interacting with robots, and robotic skins make a key contribution by enabling sensitive contact sensing and localization. This paper presents a tactile–proprioceptive sensor fusion framework for natural physical human-robot interac- tion. Tactile cues from pneumatic skin pads serve as con- tact indicators that bypass the ambiguity between frictional residues and applied external forces, enabling highly sensitive contact detection without explicit friction identification. We fuse these cues with motor-current–based proprioception to reconstruct multi-axis contact forces on the robot surface. To maintain accuracy during motion, we employ a temporal convolutional network (TCN) to mitigate friction hysteresis during stick–slip transitions, reducing uncertainty at contact onset and yielding smooth, responsive guidance. We validate the approach on a skin-integrated robot arm: (i) multi-axis forces are reconstructed in stationary contacts, and (ii) simultaneous force estimation and kinesthetic teaching are demonstrated. Results indicate improved sensitivity and responsiveness across diverse contact conditions compared with tactile-only and proprioceptive-only baselines, supporting tactile–proprioceptive fusion as a reliable pathway to safe, intuitive physical hu- man–robot interaction.