To address the issue that orbital maneuvers of low Earth Orbit (LEO) satellites interfere with the accuracy of atmospheric density retrieval and thus degrade orbit prediction performance, this study develops a maneuverdetection method based on orbit data for atmosphericdensity correction. Satellites with suitable configurations and stable ballistic coefficients are selected as reference satellites, and their precise orbits are used to invert the sphericalharmonic coefficients of the Tc and Tx parameters in the Jacchia71 atmospheric density model, thereby correcting systematic errors in the density model. In order to eliminate the contamination of maneuverperiod data on the correction process, an automatic maneuveridentification and datarejection algorithm based on differential mean semimajor axis is proposed. Validation using oneyear (2023) of measured data from Swarm satellites shows that the method can effectively detect maneuver events: for precise orbit data, the detection time resolution reaches the second level; for TwoLine Element (TLE) data, the average error of the detected maneuver midtime is less than 10 minutes, and the average error of the semimajor axis change is about 50 meters. After removing maneuversegment data, the ballisticcoefficient solutions of the reference satellites become more reasonable, with the proportion of outliers reduced by more than 90 %, significantly improving the stability and reliability of the atmosphericdensity correction coefficients. Consequently, the method contributes to enhancing the orbit prediction accuracy of LEO target satellites.
Liu et al. (Sun,) studied this question.