Research On GPS Precise Point Positioning And Atmospheric Delay Correction

GPS Precise Point Positioning (PPP) is a recently developed technique for GPS positioning, which makes use of single GPS receiver with precise ephemeris and satellite clock offset information from IGS, to get site coordinate accuracy of dm-cm level. PPP has proven to be effective and has received widespread acceptance in real-time navigation, precise time transfer, survey of land resource and aerial triangulation etc. This thesis focuses on PPP data processing, introducing its function model and statistic model, carrying out static and kinematic position solution with parameter estimation algorithms after error corrections.Three improvements of TurboEdit algorithm in GPS data preparation are presented in this thesis. In the improved algorithm, the calculating of wide-lane ambiguity and its variance takes into account different arc sections, the length of arc section and gross errors. Both combination observations determine whether the current epoch is a gross error or cycle slip by analyzing three successive epochs. A new fitting model based on sliding window is put forward, and the model is more suitable for the fact that Geometry-Free combination observations constitute a smooth time series. The secondary processing of divided Geometry-free arc sections further rejects gross error. Experimental results show that the new algorithm can detect and fix cycle slips of equal-distance, equal-cycle or continuing happened conditions. The new algorithm can also reject successfully successive gross errors and those gross errors with less obvious performance.This thesis introduces IGS's new ANTEX (antenna exchange format) file igs05.atx as well as the algorithm to get correction of both site/satellite coordinates and signal propagation range bases on this file. As for satellite clock offset information of 15min, 30s or 5s interval from IGS, three common used interpolation methods are compared and experimental results show the three methods have similar effects while dealing with IGS precise satellite clock offsets. Considering clock offset variation characteristics, a new least square fitting model based on Fourier transformation and energy spectrum analysis is presented, which overcomes Runge phenomenon of high order polynomial interpolation. Experimental results indicate that this model solves two key problems while traditional algebraic polynomial models can not: first it performs the same accuracy at any time, at any sample rate; then the interpolation results are stable. This thesis presents four algorithms to get relative effect and experimental results show that we can easily get relative effect with GPS broadcast ephemeris. Troposphere delay is a troublesome problem in GPS data processing. The processing procedure of VMF1 model, which is recommended as the most accurate by IGS Workshop 2008, is presented in this thesis.For the number of involved equations and parameters waiting estimation is too huge, sometimes we even need kinematic positioning, the general Least Square algorithm is not suitable for PPP. This thesis presents Recursive Least Square and Extended Kalman Filtering, and the developed PPP module based on the two filters. Experimental results show that the accuracies of serial epoch positioning can be better than 50cm in X , Y ,Z component after initialization of about 15 minutes for PPP based on smoothed pseudorange. While the corresponding accuracies for PPP based on un-differenced dual-frequency pseudorange and carrier phase observations are cm level. In kinematic mode, the accuracies locate between the former two.With the developed PPP module, this thesis analyses positioning accuracies affected by gross error and different intervals of satellite clock offset. The results evidently prove the positive significance for PPP from gross error rejection and higher interval of satellite clock offset information.