LEO satellite clock determination in real-time with predicted orbits introduced

Low Earth Orbit (LEO) satellites have been discussed to augment the traditional GNSS-based Positioning Navigation and Timing (PNT) service in real-time, in which the high-precision real-time LEO satellite clock products are the prerequisite. As the complicated systematic effects contained in the LEO satellite clock estimates hamper high-precision mid- to long-term clock prediction, a typical and efficient method to obtain high-precision real-time LEO satellite clocks is Kalman-filter-based clock estimation with short-term prediction. The strong correlation between the LEO satellite clock and the radial orbital component, however, leads to poorer clock precision than needed. In this contribution, reduced-dynamic LEO satellite orbits are first estimated in batch least-squares adjustment with high accuracy in near real-time. The short-term predicted orbits are introduced and constrained during the Kalman-filter-based clock estimation process. The variance-covariance matrix of the introduced orbital errors is carefully considered and tested for different sets of values in the radial, along-track and cross-track directions. One week of GPS data from the Sentinel-3B onboard receiver in 2018 were used for the purpose of validation. When introducing high-accuracy predicted orbits at the first 5 min, i.e., with an accuracy of 3.33, 1.78 and 2.03 cm in the along-track, cross-track and radial direction, respectively, the precision of the estimated clocks can be improved from 0.268 ns to 0.233 ns, with an improvement of 13.06%. Moreover, the Signal-In-Space Range Error (SISRE) of the LEO satellite to the Earth can be improved from about 9.59 to 7.62 cm after introducing the predicted orbits. The results have demonstrated that the proposed method helps to improve the precision of the real-time LEO satellite clock estimates.