Getting Close: Multi-Body Tracking Enables Camera-based Close Proximity Robotic Manipulation for On-Orbit Servicing

Anne Elisabeth Reichert, Maximilian Ulmer, Rudolph Triebel, Maximilian Durner

Keywords: Pose Estimation, Tracking, Manipulation, On-Orbit Servicing, Satellites

Abstract: Reliable 6D pose tracking of uncooperative targets is a prerequisite for autonomous on-orbit servicing and space debris removal. While terrestrial trackers have reached high levels of maturity, the orbital environment introduces unique challenges: extreme high-contrast lighting, grayscale sensor data, and highly symmetric geometries. This work extends the state-of-the-art M3T framework to address these specific failure modes. Previously, we introduced a linear motion prior to constrain the optimization space and a Bayesian fusion approach that integrates learned segmentation masks with classical intensity histograms. In this work, we present our ongoing research on multi-object configuration to address the challenges of close-range satellite images. We validate our method using synthetic path-tracing data and real-world hardware-in-the-loop trajectories, demonstrating higher resilience compared to classical region-based baselines.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 12