Enhancing Generalizable 6D Pose Tracking of an In-Hand Object with Tactile Sensing

TL;DR

TEG-Track is a novel tactile-enhanced system that improves 6D pose tracking of objects in hand by integrating tactile signals with visual data, enabling better manipulation of unseen objects.

Contribution

We introduce TEG-Track, a system that combines tactile sensing with visual tracking to improve 6D pose estimation of unseen objects in real-time.

Findings

TEG-Track outperforms existing visual-only trackers in real-world tests.

The system effectively detects slippage and adjusts pose estimates accordingly.

Our dataset facilitates future research in tactile-visual pose tracking.

Abstract

When manipulating an object to accomplish complex tasks, humans rely on both vision and touch to keep track of the object's 6D pose. However, most existing object pose tracking systems in robotics rely exclusively on visual signals, which hinder a robot's ability to manipulate objects effectively. To address this limitation, we introduce TEG-Track, a tactile-enhanced 6D pose tracking system that can track previously unseen objects held in hand. From consecutive tactile signals, TEG-Track optimizes object velocities from marker flows when slippage does not occur, or regresses velocities using a slippage estimation network when slippage is detected. The estimated object velocities are integrated into a geometric-kinematic optimization scheme to enhance existing visual pose trackers. To evaluate our method and to facilitate future research, we construct a real-world dataset for visual-tactile in-hand object pose tracking. Experimental results demonstrate that TEG-Track consistently enhances state-of-the-art generalizable 6D pose trackers in synthetic and real-world scenarios. Our code and dataset are available at https://github.com/leolyliu/TEG-Track.