Maneuver Detection Method Based on Probability Distribution Fitting of the Prediction Error

This paper presents a method for detecting historical orbital maneuvers from satellite two-line element sets. Maneuvers are detected by comparing a published state of a specific orbital parameter with a prediction state and analyzing the prediction error between them. Probability distributions of the prediction error associated with different prediction times are fitted first using sample data generated from the satellite two-line element set. The threshold for detecting outliers in the published time series of the orbital parameter is derived based on the fitting result. Subsequently, the method for deriving historical maneuvering information is designed, which ensures that a maneuver is identified by multiple consecutive outliers instead of an isolated outlier in the published time series, thereby eliminating noise interference. Maneuver detection results on typical active satellites show that the proposed method can detect historical maneuvers accurately while maintaining a low false detection rate. When the number of two-line elements for a satellite is insufficient to obtain a good detection result, a preliminary analysis proves that it is feasible to improve the detection rate by borrowing data from a nearby satellite, as long as the two satellites have similar probability distributions of the prediction error.