TactAlign: Human-to-Robot Policy Transfer via Tactile Alignment

TL;DR

TactAlign is a novel method for transferring tactile-based policies from humans to robots with different embodiments, using a shared latent space and pseudo-pairs, enabling efficient and generalizable cross-embodiment policy transfer.

Contribution

It introduces TactAlign, a cross-embodiment tactile alignment technique that does not require paired data or identical sensors, improving scalability and generality in human-to-robot tactile policy transfer.

Findings

Improves policy transfer across multiple contact-rich tasks.

Generalizes to unseen objects and tasks with minimal human data.

Enables zero-shot transfer on highly dexterous tasks.

Abstract

Human demonstrations collected by wearable devices (e.g., tactile gloves) provide fast and dexterous supervision for policy learning, and are guided by rich, natural tactile feedback. However, a key challenge is how to transfer human-collected tactile signals to robots despite the differences in sensing modalities and embodiment. Existing human-to-robot (H2R) approaches that incorporate touch often assume identical tactile sensors, require paired data, and involve little to no embodiment gap between human demonstrator and the robots, limiting scalability and generality. We propose TactAlign, a cross-embodiment tactile alignment method that transfers human-collected tactile signals to a robot with different embodiment. TactAlign transforms human and robot tactile observations into a shared latent representation using a rectified flow, without paired datasets, manual labels, or privileged information. Our method enables low-cost latent transport guided by hand-object interaction-derived pseudo-pairs. We demonstrate that TactAlign improves H2R policy transfer across multiple contact-rich tasks (pivoting, insertion, lid closing), generalizes to unseen objects and tasks with human data (less than 5 minutes), and enables zero-shot H2R transfer on a highly dexterous tasks (light bulb screwing).