Detection of Satellite Maneuvers in Earth and Cislunar Orbits using Adaptive CuSum Methods

Abstract

The tracking of satellite and resident space object (RSO) trajectories is essential to both space situational awareness and celestial navigation. As such, detection of maneuvers and other orbital anomalies of RSOs constitutes a challenge to be addressed. This work discusses the application of an adaptive automated multiple CuSum algorithm on geocentric and cislunar satellites to detect impulsive maneuvers. Historically both data and model-based approaches have been used to address this problem. These methods rely on the deviation of each data point to some predicted value and rely on parameter tuning. This study leverages and further develops the previous work by introducing a statistical perspective, the CuSum algorithm. It relies on invariant orbital elements for maneuver detection. Real data from both geocentric and cislunar satellites are used to test the methodology. Using Two Line Elements or TLE data for six Low Earth Orbit RSOs, catch rates of over 96% within a few data points after their occurence are achievable, with an occurence of only 1 false alarm per 400 days. For satellites and ephemeris real data from two flown cislunar missions the algorithm achieves catch rates of over 89 % and false alarm rates of over 50 days per false alarm. Adjusting the threshold provides a trade off in catch and false alarm rates.