Algorithm Of Extraction And Regional Modeling Of Atmospheric Error Based On Un-differenced And Un-combined Model

Accuracy, instantaneity, and flexibility are the main objectives and characteristics of GNSS technology. Using a single receiver for fast positioning is important in the future opportunities and challenges. Aiming to the problems about how to use observational data from reference stations to provide accurate priori information of atmospheric errors for PPP users within region and enhance the efficiency of the positioning, this paper carried out the work in the following aspects:On the basis of analysis on function models and stochastic models in three commonly used precise point positioning models, this paper discusses and researches of the various error processing strategy and detailed setup of the filter when using un-differential and un-combined model to extract atmospheric error of stations. Through comparison and analysis of ionosphere-free combined model, the comparative advantages of atmospheric delay information extraction using un-differential and un-combined model were obtained.A regional modelling of atmospheric delay method with elevation difference in consideration was presented. This method extracted zenith tropospheric delay of stations using un-differential and un-combined precise point positioning model. Then tropospheric delay was divided using UNB3m empirical model based on the spatiotemporal characteristics of zenith tropospheric delay, according to the relationship between the troposphere delay and station elevation, and the two components correlated with elevation weakly and strongly were obtained. At last, the user station’s zenith tropospheric delay was obtained by interpolation modelling of the two components separately within region. The results show that:the root mean square errors of HTIM were within 0.020m.A modified linear combination method based on ionospheric pierce point of satellites was proposed using ionosphere delay of stations in slant direction of each satellite extracted by un-differential and un-combined precise point positioning model and considering thin layer hypothesis theory of ionosphere delay. Firstly, the impact of differential code bias on the satellites was eliminated by the DCB files released by IGS. Then the ionosphere delay was interpolated regionally using three-dimensional coordinates of pierce point. Experimental results show that either in the low-latitude or high-latitude areas, this method can use rather sparse reference stations for regional modelling. And the obtained interpolation results can provide priori ionospheric delay with the accuracy within 0.1m for the rover station compared with the true value after removing the system bias introduced by the interpolation using differential between satellites.At the same time, an un-combined single-difference (SD) between satellites method on the rover end based on the results of interpolation was proposed according to the characteristics of interpolation results mentioned above. The SD Kalman filter was built to eliminate differential code bias on the rover receiver and system bias by the interpolation of ionosphere delay and the PPP rapid positioning of the rover station was achieved using the priori values of ionosphere delay and troposphere delay obtained by regional modelling. For rover users, the convergence time of positioning was reduced greatly while the precision of positioning results was ensured. The un-combined single-difference between satellites precise point positioning method using atmosphere information can make sure that the convergence time is less than 30 minutes in the three directions of north, east and up, and reduced by 10% when compared with the un-combined un-difference method.Based on the theoretical research, the algorithms of extraction and regional modeling of atmospheric error mentioned in the paper are added into the service module of regional augmented PPP and integrity monitoring in the GNSS ground-based augmentation systems EarthNet. The combination of the theoretical research results and the system improved the value of engineering application of the related algorithms.