Geometric Approaches to 6-DoF Pose Estimation for Autonomous Spacecraft Docking

Precise six-degree-of-freedom (6-DoF) pose estimation is essential for autonomous spacecraft docking, where even minor errors can compromise mission safety. Classical attitude determination algorithms such as QUEST, FOAM, and ESOQ are effective for rotational estimation, but are insufficient for full pose recovery due to the absence of translation estimation. The evolution of Perspective-n-Point (PnP) formulations has enabled complete pose estimation through geometric correspondences; however, these methods rely heavily on ideal imaging conditions, which are rarely available in space environments, especially under thermal imaging. This paper presents a comprehensive review of geometry-driven 6-DoF pose estimation methods for autonomous docking, emphasizing purely geometric approaches without reliance on deep learning. A baseline framework grounded in classical attitude solvers and PnP formulations is discussed and existing research gaps—particularly under thermal sensing constraints—are identified. The study establishes a solid foundation for future real-time, interpretable, and thermally adaptive spacecraft docking systems.