GPS Precise Point Positioning Algorithm Using One-Way Phase Observations

Precise point positioning (PPP) uses un-differenced dual frequency pseudorange and carrier phrase observations, along with IGS precise orbit and satellite clocks, for stand-alone precise geodetic point positioning (static or kinematic) with centimeter precision. The first step for GPS Precise Point Positioning (PPP) with one-way phase model is to obtain information about IGS precise orbit and satellite clock. The second step is to make use of observations of dual-frequency pseudorange and carrier measurements received from one receiver. Some errors coming from station, satellite and transmit path of signal will affect the positioning result. Thus, error corrections are the necessary procedure. One the other hand, Kalman filter algorithm is an effective way in GPS kinematic data process to improve the accuracy of point positioning. The mathematical models and algorithms of GPS un-differenced precise point positioning are investigated systematically in this paper. The un-differenced data process and the usage of Kalman filter algorithm in the PPP are primarily analyzed. As a result, un-differenced PPP software is developed. This paper mainly focuses on the following three parts:1. Study on cycle-slip processing method of phase difference between adjacent epochs, then develop program to validate this method. The result indicates that the cycle-slip can be rapidly and accurately detected and repaired by this method. It has high practicality and efficiency, especially for kinematic GPS positioning;2. Discussion of some problems related to the constructions of the state equation and observation equation in Kalman filter model, and respectively analyses linearization of the observation equation, selection of the state covariance matrix value and initialization of the filter parameters;3. Development of the PPP software with un-differenced model in Visual C++ circumstance, and use the software to analyze the affection to the result by giving different initial covariance and to compare the accuracy of result by using different sampling interval data. The experiment indicates that appropriate initial covariance can make filter converge fast, empirical initial covariance selected by this paper can make the filter has a fast convergent speed, but it does not affect the accuracy of result after the filter become steady. In the end, compares the calculated results by the data of different sampling interval, it proves that the result calculated by the data of high sampling interval has a higher accuracy than the data of low sampling interval.