Reliable Atmospheric Correction Generation Method for Out-of-Network PPP-RTK Users

Abstract

Precise point positioning–real-time kinematic (PPP-RTK) technology can achieve accurate regional positioning with precise atmospheric corrections. However, traditional PPP-RTK algorithm research has focused on users in a reference station network. For out-of-network users, the extrapolated corrections typically contain significant errors, because it is difficult to model atmospheric delays within an area without sufficient reference stations. Poor-quality atmospheric corrections can significantly affect integer ambiguity resolution (IAR). We propose a reliable atmospheric correction generation method for out-of-network PPP-RTK users to maintain a continuous, ambiguity-fixed state in an inaccurately corrected observation environment. When users find that they have moved out of the reference station network, integer ambiguities acquired from the last epoch are used to estimate atmospheric corrections using an inversion calculation method. Due to the short epoch interval employed by PPP-RTK users, the estimated atmospheric corrections can properly correct delays in the observations of the current epoch. Before the users are forced to use the ionosphere-float model, they can employ the proposed method to continue utilizing the ionosphere-fixed model, maintaining the instantaneous ambiguity-fixed state for an extended period. This approach circumvents the need to directly adopt the ionosphere-float model, which introduces ionosphere parameters and necessitates time for them to converge. In experiments, both the integer ambiguity resolution success rate and correct rate demonstrated the superiority of the proposed method. In a kinematic PPP–ambiguity resolution (AR) experiment, using the estimated corrections, the mean success fix rates were within the range 31%–74%, and 93% of the horizontal errors were controlled to within 3 cm using the estimated corrections. This was significantly more favorable than the results obtained using traditional extrapolated corrections. It was confirmed that the proposed method possesses a much stronger ability to provide continuous ambiguity-fixed solutions and improve the accuracy of the coordinates for out-of-network users.