Prediction versus real-time orbit determination for GNSS satellites

Abstract

To serve real-time users, the IGS (International GNSS Service) provides GPS and GLONASS Ultra-rapid (IGU) orbits with an update of every 6 h. Following similar procedures, we produce Galileo and BeiDou predicted orbits. Comparison with precise orbits from the German Research Centre for Geosciences (GFZ) and Satellite Laser Ranging (SLR) residuals show that the quality of Galileo and BeiDou 6-h predicted orbits decreases more rapidly than that for GPS satellites. Particularly, the performance of BeiDou IGSO and MEO 6-h predicted orbits is 5–6 times worse than the corresponding estimated orbits when satellites are in the eclipse seasons. An insufficient number and distribution of tracking stations, as well as an imperfect solar radiation pressure (SRP) model, limit the quality of Galileo and BeiDou orbit products. Rather than long time prediction, real-time orbit determination by means of a square root information filter (SRIF) produces precise orbits every epoch. By setting variable processing noise on SRP parameters, the filter has the capability of accommodating satellite maneuvers and attitude switches automatically. An epoch-wise ambiguity resolution procedure is introduced to estimate better real-time orbit products. Results show that the real-time estimated orbits are in general better than the 6-h predicted orbits if sufficient observations are available after real-time data preprocessing. On average, 3D RMS values of the real-time estimated orbits reduce by about 30%, 60% and 40% over the 6 h predicted orbits for GPS, BeiDou IGSO and BeiDou MEO eclipsing satellites, respectively. Galileo satellites did not enter into the eclipse season during the experimental period, the standard derivation (STD) of SLR residuals for the real-time estimated orbits are almost the same as for the post-processed orbits.