Method Of The Satellite Clock Error Interpolation In The Post-Processing Of PPP

GPS Precise Point Positioning (PPP) is a new GPS technique which developed in recent years. It mainly uses precise ephemeris and satellite clock errors provided by IGS. It also uses un-differenced phase and pseudorange which collected by single receiver as the main observations. Decimeter-level accuracy can be achieved even centimeter-level. It can use a single receiver to determine the three-dimensional coordinates in the International Terrestrial Reference Frame (ITRF) for geodetic point positioning (static or kinematic) on the global scale. So PPP has broad application prospects in many areas of Geodesy. It is the hot topic of GPS presently.In PPP, GPS satellite clock error is one of the main factors of influence of positioning accuracy. For the high sensitivity and high frequency of GPS atomic clock and highly vulnerable to the external or its own factors, it is difficult to grasp the complexity and the detail changes. So how to interpolate high quality GPS satellite clock becomes one of the main technical questions. Therefore, the satellite clock error correction is a necessary means to improve its accuracy in the post-processing of the PPP.In this paper, it firstly gives an outline of the development status, principle and algorithm of PPP, and then analyzes various errors which impact the positioning accuracy and some corrective measures. The paper introduces five interpolation methods in details and has programmed these algorithms. It compares some different types of satellite clock with some examples and analyzes the pros and cons of various interpolation methods. The results show liner interpolation method is optimal for the satellite clock which clock errors'changes are in stability and spline interpolation is optimal for the satellite clock which clock jitter is big. Secondly, it discusses two different data models. It mainly compares the accuracy between the time model which is near to the interpolation point and the 24-hour time model. The result shows that we can choose the time model which is near to the interpolation point to process data, for it not only meets the accuracy requirements, but also saves time.