URE and URA for Predicted LEO satellites Orbits at different altitudes

Abstract

In recent years, low Earth orbit (LEO) satellites have been frequently discussed for their benefits in positioning and navigation services as an augmentation to the global navigation satellite systems (GNSSs). Similar to the positioning concept based on ranging to GNSS satellites, precise positioning of single-receiver users needs high-accuracy orbits and clocks of LEO satellites as a pre-condition. For real-time users, high prediction accuracies of these orbits at different latencies are needed. Unlike the satellite clocks, the GNSS orbits can be typically predicted for hours with high accuracy. LEO satellites, however, face more complicated perturbing dynamic terms due to their low altitudes. Therefore, the prediction accuracy and integrity of their orbits need to be addressed. In this study, using real data of three test LEO satellites GRACE C, Sentinel-1A and Sentinel-3B of different altitudes, various reduced-dynamic prediction strategies are assessed, with the appropriate methods selected for different prediction times up to 6 h. The global-averaged orbital user range errors (OUREs) are shown to be altitude-related. For the 700–800 km Sentinel satellites and 500 km GRACE satellite, the RMS of the OUREs is at sub-dm and dm-level for the prediction time of 1 h, respectively, and around 0.2 m and 0.6 m at the prediction time of 6 h, respectively. For integrity purposes, the worst-location OURE are calculated for the predicted orbits using a proposed algorithm considering the Earth as an Ellipsoid, not a sphere as usually done for the GNSS satellites. The orbital user range accuracy (OURA) is then evaluated for different prediction periods, having a time-dependent model proposed to compute the overbounding OURA at any prediction time within 6 h. With an integrity risk of 10 5, using hourly quadratic polynomials as the time-dependent model, the overbounding OURA is around 0.1 m at the prediction of 1 h, and at the sub-meter level for the prediction of 6 h for the Sentinel satellites. 2022 COSPAR. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/