InvariantCloud: A Globally Invariant, Uniquely Indexed Point Cloud Framework for Robust 6-DoF Tactile Pose Tracking

2026-05-24

AI Summary

This work addresses the challenges of cumulative drift and inaccurate yaw-angle (rotation about the Z-axis) estimation in vision-based tactile sensors for 6-degree-of-freedom pose tracking. To overcome these limitations, the authors propose a point cloud registration framework leveraging the global invariance of surface marker constellations. This approach introduces, for the first time, a globally invariant point cloud representation combined with a one-shot registration strategy to achieve drift-free and robust pose tracking. By incorporating uniquely indexed global invariant features and explicit modeling of surface marker constellations, the method significantly enhances yaw-angle estimation accuracy and relocalization repeatability. Experimental results demonstrate that, in long-sequence manipulation tasks, the proposed method outperforms existing techniques in both yaw-angle tracking and relocalization performance, confirming its high precision and strong robustness.

Abstract

Recent advances in imitation learning and vision-language models highlight the need for high-fidelity tactile perception, with 6-DoF tactile object pose estimation providing a crucial foundation for precise robotic manipulation. We introduce InvariantCloud, a 6-DoF pose estimation framework that leverages the global invariance of surface marker constellations on vision-based tactile sensors. In contrast to recent approaches, our one-shot globally invariant point cloud registration suppresses cumulative drift and overcomes long-standing limitations in accurately estimating yaw (Z-axis) rotation. Experimental verifications show that InvariantCloud achieves superior yaw tracking accuracy and re-localization repeatability compared to existing benchmarks, demonstrating its precision and robustness in long-sequence manipulation tasks.

Problem

• 6-DoF pose estimation
• tactile sensing
• yaw rotation
• point cloud registration
• pose tracking

Innovation

• globally invariant
• point cloud registration
• 6-DoF tactile pose estimation
• yaw accuracy
• tactile sensing