High-Contrast Detection of Co-orbital Satellites at Geosynchronous Orbit

The dramatic increase in co-orbital satellites in geosynchronous equatorial orbit (GEO) results from significant technological advances in rendezvous and proximity operations (RPO) over the past couple of decades. RPO enables spacecraft to maneuver close to another space object and provides many functions such as on-orbit refueling, assembly, debris removal, and inspection. There are also defense-related mission sets that include on-orbit intelligence, surveillance, and reconnaissance (ISR) options critical for monitoring at GEO due to the prevalence of telecommunication satellites that the military relies on for operations worldwide. RPO also provides a potential co-orbital anti-satellite adversarial offensive capacity, which makes it essential to maintain detection capability of these types of on-orbit systems. Consequently, RPO's in GEO pose a unique space domain awareness challenge for ground-based sensors that attempt to track and identify space assets. The primary space domain awareness (SDA) tracking systems used to monitor assets in GEO are large optical ground-based telescopes. Imaging and detecting RPO engagements in GEO are problematic for two main reasons. First, the satellites in GEO are typically telecommunication satellites that require a sizeable structural framework to hold a bulk of analog equipment and fuel to operate for the length of their assignment. In comparison, an on-orbit RPO satellite would likely be smaller in size, depending on its mission. The difference in comparative size and corresponding optical cross-sectional area of on-orbit RPO satellites make them much fainter objects to detect. Second, RPO operations inherently bring spacecraft close to one another. RPO close-proximity capabilities, satellites could maneuver inside of a 1 km separation or less, making it challenging to distinguish from one another with current ground-based SDA optical imaging capabilities. With these two factors, the significant difference in visual magnitudes and small separation distances bury the fainter RPO satellite in the optical signature profile of the larger GEO satellite. In this scenario, the images of each of the satellites can appear merged even with an adaptive-optics (AO) enabled telescope due to limitations of the optical system's resolving power, sensor pixel size, and the uncompensated effects of imaging through turbulence. This dissertation explores an objective task-based assessment approach to detecting a co-orbital RPO satellite near a larger GEO satellite and is addressed in three main parts. First, the ideal linear discriminant, otherwise known as the Hotelling observer, is derived and applied to AO compensated images in a wave-optics simulation of a ground-based space domain awareness telescope. Additionally, the second part focuses on using real-time data updates to the prewhitening filter of a linear discriminant to adapt to changes in atmospheric seeing conditions to improve robustness. Finally, we explore the fusing of decisions from multiple telescopes to enhance detection capabilities from a network of decentralized SDA sensor systems. These techniques will be compared with the matched filter approach to provide a baseline using receiver operating characteristic curves (ROC) in the statistical decision theory framework.Dissertation not available (per author's request