A Cost-Effective Crowdsourced Q4DIM Method for Rapid PPP Implementation in Wide Areas

As Global Navigation Satellite System (GNSS) positioning technology rapidly advances, Precise Point Positioning (PPP) has found widespread application in the mass market, particularly for vehicle navigation. PPP enhanced by atmospheric corrections was proven to be an effective rapid high-precision positioning method for wide-area users. The Quasi‐4‐Dimension Ionospheric Modeling (Q4DIM) is a flexible atmospheric model that enables enhanced PPP positioning from wide-area to regional applications. It divides the slant ionospheric delays (SIDs) from a station network into various clusters based on the latitude and longitude of the Ionosphere Piercing Point (IPP), satellite elevation, and satellite azimuth. These clusters are used for the correction of atmospheric errors in PPP and ionosphere monitoring. As devices capable of GNSS positioning become increasingly available in the mass market, effectively utilizing these observations could significantly reduce costs and broaden the range of rapid PPP services. In this contribution, we developed a crowdsourcing Q4DIM approach, where users upload SIDs verified for integrity to the cloud server, which classifies and stores the data based on accuracy, location, and the level of services utilized. Then, the cloud server constructs and disseminates diversity Q4DIM maps according to the different level attributes of SIDs. Finally, the users utilize the updated Q4DIM maps to achieve faster and more precise positioning. 144 sets of control experiments are conducted with observations from European Continuously Operating Reference Stations (CORS). Stations with an average inter-station distance of about 200 km are chosen as reference stations that are used for extracting the original Q4DIM map. The remaining stations are established as dynamically estimated crowdsourced stations for extracting the crowdsourced Q4DIM map. Results show that the performance of PPP enhanced by the crowdsourced Q4DIM map is significantly improved than those of the original Q4DIM map. The positioning error series of the original solution converges within 9 epochs to within 10 cm in the horizontal direction and 20 cm in the vertical direction, while the positioning error series of the crowdsourced solution reaches 2.3 cm in the horizontal direction and 7.6 cm in the vertical direction in 2 epochs. Compared to the original solution, the positioning accuracy of the new method improved by 48.2% in the horizontal direction and 41.2% in the vertical direction.