Research On Several Key Issues Of GNSS Satellites Orbit Determination Based On Realtime Data Stream

At present,GNSS real-time precise positioning service has been widely used in social production and scientific research fields,such as mapping and remote sensing,automatic driving,and disaster monitoring.The real-time precise orbit products of GNSS satellites are one of the important basements to provide global real-time precise positioning service.However,the accuracy and reliability of the predicted orbits depend on the stability of the satellite motion and force model.It is difficult to maintain the accuracy and reliability of the predicted orbits when the satellites are abnormal.With the steady improvement in the quality of real-time observation,there is an urgent need to develop the precise orbit determination(POD)method based on filtering and real-time data stream to enhance the capability of real-time positioning service.Based on GNSS real-time POD,this dissertation focuses on the real-time data stream anomalies and navigation satellite dynamics model anomalies,and carried out studies on the missing observation and uneven station distribution of the IGS real-time data stream,quasireal-time maneuver detection,and POD using adaptive filtering.The validity of the proposed algorithms and models has been verified by testing on real data.The main research work and contributions of this paper are as follows:1)Taking the IGS real-time data stream as an example,this paper address issues about uneven station distribution and missing observation of the real-time data stream.To improve the accuracy considering real-time data stream anomalies,a real-time POD based on filtering constrained by ultra-rapid orbits has been proposed.Statistics of results show that the impact of uneven station distribution on GPS and Galileo satellite orbit RMS is only0.3?0.5cm in along and normal direction.The missing observation is the main factor affecting the orbit accuracy.Constrained by ultra-rapid orbits,the average 3D RMSs of GPS and Galileo orbits were improved by 12% and 34%,respectively.2)For the orbit maneuver,a quasi-real-time maneuver detection method based on residual of carrier phase triple-differenced observation is proposed.The experimental results show that the accuracy of the maneuver end time detected by this method is better than 5min,which can effectively reduce the ephemeris unavailability time caused by the orbit maneuver.3)For the meter-level or decimeter-level satellite dynamics model anomalies,new adaptive filtering for real-time POD is proposed,which is based on the inconsistency of orbit paraments.The real data during G05 and C11 anomalies are used to verify this method.The results show that the along maximum orbit error of the G05 is reduced from 1.3 m to0.5 m.And that of the C11 is reduced from 2.2 m to 1.5 m.The orbit 3D RMSs of the two satellites during dynamic model anomalies were improved by 40% to 70%.4)Based on the FUSING software,POD modules considering real-time data stream anomalies and satellite dynamics model anomalies have been developed to achieve robust processing of real-time POD.The IGS real-time observation data from DOY 029 to DOY042 in 2022 were used for real-time POD validation.Compared with the IGS and CODE final products,the average 3D RMSs of orbits are 5.1 cm and 7.1 cm for GPS and Galileo satellites,respectively,which are better than the 3?9 hours predicted orbit of IGU for GPS satellites and the 1?2 hours predicted orbit of WHU products for Galileo satellites.And the average STDs of satellite clocks estimated together the satellite orbits are 0.036 ns and0.034 ns for GPS and Galileo satellites,respectively.