Built Different: Tactile Perception to Overcome Cross-Embodiment Capability Differences in Collaborative Manipulation

TL;DR

This paper introduces a tactile sensing approach combined with behavior cloning to enable robots with different embodiments, especially those lacking torque sensing, to perform collaborative manipulation tasks effectively.

Contribution

It presents a novel method for transferring manipulation policies across different robot embodiments using tactile sensors and behavior cloning, including robots without torque sensing capabilities.

Findings

Successful policy transfer across four different tactile-enabled robots

Decomposed tactile shear-field representation improves success rates

Pre-trained encoder enhances cross-embodiment performance

Abstract

Tactile sensing is a widely-studied means of implicit communication between robot and human. In this paper, we investigate how tactile sensing can help bridge differences between robotic embodiments in the context of collaborative manipulation. For a robot, learning and executing force-rich collaboration require compliance to human interaction. While compliance is often achieved with admittance control, many commercial robots lack the joint torque monitoring needed for such control. To address this challenge, we present an approach that uses tactile sensors and behavior cloning to transfer policies from robots with these capabilities to those without. We train a single policy that demonstrates positive transfer across embodiments, including robots without torque sensing. We demonstrate this positive transfer on four different tactile-enabled embodiments using the same policy trained on force-controlled robot data. Across multiple proposed metrics, the best performance came from a decomposed tactile shear-field representation combined with a pre-trained encoder, which improved success rates over alternative representations.