Sensitivity and comparison of orbital state-based maneuver detection approaches

Detecting and characterizing orbital maneuvers is crucial for many tasks related to Space Situational Awareness, such as space catalogue maintenance, or identification of anomalous behaviors and associated threats. Many maneuver detection methods are based on the analysis of the time history of orbital data, i.e., orbital ephemeris or Two Line Elements sets. In general, these methods identify manoeuvres as anomalies in the orbital time history, but a basic distinction can be made between derivative-based and propagation-based approaches, depending on the use of data fitting or orbital propagation in the process, respectively. Hybrid methods also exist, trying to merge the advantages of the two approaches. The performance of these methods is affected by many factors, such as input data uncertainty, data update frequency, selection of tuning parameters, as well as uncertainties of orbital propagation and type and magnitude of the orbital maneuvers to be detected. In this context, this paper discusses the implementation and the performance analysis of three maneuver detection methods, specifically two derivative-based methods and a propagation-based one. The analysis is first carried out using simulated test cases, varying orbital maneuvers’ magnitude and tuning parameters. Then, an application of the three methods for two real test cases is presented and discussed.