SBIR-STTR Award

The current Space Surveillance Network (SSN) is projected to soon be unable to track all manmade Low Earth Orbit (LEO) objects at the current rate of observation. Very large constellations of satellites will exponentially grow the number of LEO objects, and simply adding more sensors to the SSN to keep pace with the proliferation in LEO is a cost-prohibitive proposition. This research proposes to address the proliferation of LEO constellations/objects using novel estimation algorithms without the need to expand or improve the current SSN configuration. In particular, the work focuses a novel estimation algorithm with nearly linear-time complexity that incorporates multi-fidelity orbit propagation, a combination particle and Gaussian sum filter update, and built-in, statistically optimal data association.

The current Space Surveillance Network (SSN) is projected to soon be unable to track all manmade Low Earth Orbit (LEO) objects. Unprecedentedly large constellations of satellites will exponentially grow the number of LEO objects, and simply adding more sensors to the SSN to keep pace with the proliferation in LEO is a cost-prohibitive proposition. This research proposes to address the proliferation of LEO constellations/objects using novel estimation algorithms without the need to expand or improve the current SSN configuration. In particular, the work focuses a novel estimation algorithm with nearly linear-time complexity that incorporates multi-fidelity, GPU-based orbit propagation, a hybrid Gaussian mixture sampled particle filter with adaptive domain partitioning, and an adaptive, statistically optimal multi-hypothesis filter that can handle massive constellations, nearly spaced objects, and slowly maneuvering satellites.