New manoeuvre detection method based on historical orbital data for low Earth orbit satellites

A new method of deriving satellite manoeuvring information from historical two-line element (TLE) data is proposed. The historical manoeuvres of a satellite are detected by identifying abnormal data segments in the TLE derived time series of selected orbital parameters (semi-major axis and inclination). Firstly, moving window approach is used to divide the time series into a series of equal-length data segments. Secondly, two anomaly indexes are introduced to measure the anomaly degree of the semi-major axis segment and the inclination segment with respect to the propagated states. Finally, the corresponding detection thresholds are separately derived by analysing the anomaly indexes of the two types of data segments indicative of orbit manoeuvre. When the anomaly index of a semi-major axis segment or inclination segment exceeds the corresponding threshold, a specific type of orbit manoeuvre with specific magnitude is declared. The manoeuvre detection results of two low Earth orbiting satellites indicate that the proposed method can efficiently eliminate data noise interference and accurately detect historical manoeuvres. Furthermore, manoeuvres with designated magnitudes can be exclusively detected by adjusting the user-specified arguments in the threshold expression levels.