OSMO: Open-Source Tactile Glove for Human-to-Robot Skill Transfer

TL;DR

OSMO addresses the gap in tactile feedback for manipulation by introducing an open-source tactile glove with 12 magnetic sensors that capture both shear and normal forces. It enables training robot policies entirely from human tactile demonstrations using a diffusion-policy framework, preserving tactile information during deployment without robot data. The approach improves performance on a contact-rich wiping task, outperforming vision-based baselines and reducing contact-related failures. By releasing full hardware and software, OSMO lowers barriers to collecting tactile demonstrations and facilitates broader adoption of human-to-robot skill transfer.

Abstract

Human video demonstrations provide abundant training data for learning robot policies, but video alone cannot capture the rich contact signals critical for mastering manipulation. We introduce OSMO, an open-source wearable tactile glove designed for human-to-robot skill transfer. The glove features 12 three-axis tactile sensors across the fingertips and palm and is designed to be compatible with state-of-the-art hand-tracking methods for in-the-wild data collection. We demonstrate that a robot policy trained exclusively on human demonstrations collected with OSMO, without any real robot data, is capable of executing a challenging contact-rich manipulation task. By equipping both the human and the robot with the same glove, OSMO minimizes the visual and tactile embodiment gap, enabling the transfer of continuous shear and normal force feedback while avoiding the need for image inpainting or other vision-based force inference. On a real-world wiping task requiring sustained contact pressure, our tactile-aware policy achieves a 72% success rate, outperforming vision-only baselines by eliminating contact-related failure modes. We release complete hardware designs, firmware, and assembly instructions to support community adoption.

Figures (7)

• Figure 1: (A) The OSMO tactile glove for collecting in-the-wild human demonstrations provides full-hand coverage with 3-axis tactile sensors, integrates seamlessly with hand-tracking systems, and can be directly deployed on robots. (B) We demonstrate that a contact-rich wiping policy trained solely on OSMO tactile-glove human demonstrations can be directly deployed on a robot, outperforming vision-only policies. More videos, code, and design files are available at \website.
• Figure 2: The OSMO tactile glove is designed to be compatible with off-the-shelf hand-tracking devices for in-the-wild data collection. OSMO is worn on the right hand, and the device’s native hand-tracking estimates are overlaid. (A) Aria Gen 2, pick and place. (B) Quest 3, ironing. (C) Apple Vision Pro, espresso machine operation. (D) RGB video, pepper-grinder use. (E) Manus glove integration for hand-pose tracking robust to visual occlusion. The OSMO magnetic tactile sensors do not interfere with Manus finger tracking.
• Figure 3: (A) Layer-by-layer breakdown of the magnetic sensor assembly, including MuMetal shielding designed to reduce crosstalk and improve signal fidelity. (B) Wiring diagram of the sensors and microcontroller boards. Data is output via USB-C.
• Figure 4: Crosstalk characterization experiments with the glove worn by a robot hand. (A) Finger motion: the robot executes a sinusoidal finger wave. (B) Adjacent contact deformation: the soft magnetic patch on the index distal phalanx is repeatedly pressed.
• Figure 5: A) Glove layout maximizing sensing coverage while minimizing encumbrance. B) The shared glove platform minimizes the visual gap between the human and robot hand.