Study On Real-time Multi-GNSS Precise Point Positioning And Sparse Representation Of Tropospheric Delay

The completion of BeiDou global network has greatly enriched the research on GNSS multi-frequency and multi-system precision positioning.Precise Point Positioning(PPP)requires only a single receiver to achieve high-precision absolute positioning.In recent years,communication means and communication speed have developed rapidly,research on real-time multi-GNSS PPP has important application value.Real-time multi-GNSS cannot only obtain high-precision position information,but also can monitor the changes of the atmosphere in real time for GNSS meteorological research.It can combine with the dynamical model achieve real-time orbit determination of Low Earth Orbit satellites(LEOs),and can be widely used in many fields such as precise time transfer and earthquake monitoring.This paper takes the four-system(GPS,GLONASS,BDS and Galileo)GNSS real-time precise point positioning(RTPPP)as the core.Studies the influence of different combinations,PPP ambiguity resolution on the static and dynamic precise point positioning,and evaluates the accuracy of tropospheric delay through PPP,and the numerical stability of SRIF(Square Root Information Filter)filtering and Kalman filtering are compared.Based on the compressed sensing method,real-time sparse reconstruction of the Global Tropospheric Gird Data(GTGD)are achieved.In order to ensure the carrier data quality of real-time LEOs orbit determination by combining real-time multi-GNSS precision point positioning and dynamical model,a new method QCCLK,based on receiver clock error index to detect carrier quality problem,is proposed.The main research contents and innovations of this paper are as follows:(1)We discussed different filtering methods(Kalman and SRIF filter)and observations random model effect on the accuracy of convergence time and accuracy.In addition,we proved that SRIF filter has higher numerical stability than Kalman filter under the limited computer word length.The effects of different navigation system combinations on positioning accuracy,tropospheric delay accuracy,PPP convergence time,and PPPAR on positioning error are compared.(2)To reduce the real-time transmission load of global tropospheric grid data,the K-SVD based compressed sensing(CS)method is used for sparse reconstruction for the first time to sparsely reconstruct the global tropospheric grid data.The results show that sparse real-time global tropospheric grid model has higher accuracy than spherical harmonic expansion modeling.To reduce the memory and number of calculations required for the K-SVD(K-means and SVD)algorithm,the mini-batch K-SVD algorithm is proposed to speed up the calculation process.This article discusses several key problems of CS processing and application in GTGD,such as signal sparsity,reconstruction accuracy,iteration times,etc.The sparse reconstruction ability of GTGD data with dictionaries with different iteration times and sparsity is analyzed.(3)A new method of detecting the quality of carrier phase observations based on the receiver clock offset index(QCCLK)is proposed,which fully combines the dynamical model and the GNSS observations to calculate the change of the clock offset of each satellite receiver to value the quality of the carrier phase observations.(4)Use the real-time data stream product(CLK93)released by CNES(Centre National d’Etudes Spatiales)and the GRACE Follow-On satellite(Level-1 A)observation file to perform real-time precision orbit determination experiment.The results show that spaceborne real-time autonomous orbit determination accuracy of the broadcast ephemeris is about 0.50 m(3D RMS),and the real-time data stream product(CLK93)has an accuracy of about 0.10 m(3D RMS)in real-time orbit determination.