Francisca Parra, Air Force Institute of Technology; Glenn Nestell, Air Force Institute of Technology; Rachel Oliver, Air Force Institute of Technology
https://doi.org/10.64861/XPSM9600
Keywords: Optical Tracking, Space Domain Awareness (SDA), Low Earth Orbit (LEO), Medium Earth Orbit (MEO), Geostationary Orbit (GEO), Event-Based Sensors (EBC), Cooperative Active Sensing, Orbit Determination, Ground-Based Optical Systems
Abstract:
Angles-only orbit determination remains a central challenge for optical Space Domain Awareness (SDA) because optical sensors provide accurate line-of-sight measurements but do not measure range directly. As a result, the quality of an orbit solution depends not only on the sensing modality, but also on the number of available measurements, observation arc length, temporal spacing, angular-rate behavior, observer-target geometry, and orbital regime. These dependencies become especially important when comparing conventional frame-based CMOS cameras and event-based cameras (EBCs), since each sensing paradigm produces different measurement products and may favor different orbit determination approaches.
This paper experimentally assesses selected classical and modern angles-only OD methods using real optical measurements from CMOS and event-based sensing systems. The study evaluates Gauss initial orbit determination (IOD), batch least-squares estimation (LSE), and the Advanced Uni-sensor Rapid Orbit Reconstruction Algorithm and Sensing (AURORAS) concept using common, bounded observation arcs. This limited-arc framework reflects practical optical SDA conditions, where complete or long-duration observation arcs are not always available, and enables a consistent comparison of method and sensor performance under equivalent measurement conditions. Gauss IOD represents a classical minimum-data approach based on three discrete angular observations, LSE represents a multi-measurement refinement approach based on residual minimization over the selected arc, and AURORAS represents a modern rapid-reconstruction approach that exploits high-rate angular motion information available from event-based or high-temporal-resolution optical sensing.
The evaluation considers representative Low Earth Orbit (LEO), Medium Earth Orbit (MEO), and Geosynchronous Earth Orbit (GEO) cases to assess how orbital regime affects method performance. The study processes time-tagged angular measurements into orbit estimates and compares the resulting solutions using common metrics related to angular residuals, orbit reconstruction error, convergence behavior, observation arc requirements, and practical suitability for SDA operations. It examines whether camera type primarily determines angles-only OD performance, or whether measurement cadence, arc length, angular separation, apparent target motion, observing geometry, and sensor output interact to drive orbit reconstruction quality. The outcome is a clearer understanding of when conventional CMOS measurements provide sufficient information for classical or batch OD methods, when EBC-derived high-rate measurements can improve rapid reconstruction, and how these tradeoffs vary across LEO, MEO, and GEO regimes. This work provides a structured experimental comparison of Gauss, LSE, and AURORAS under a unified optical SDA framework and supports the selection and integration of classical and modern angles-only OD methods for future tracking architectures.
Date of Conference: September 15-18, 2026
Track: Space-Based Assets