Ionosphere interpolation for undifferenced and uncombined PPP-RTK in Hong Kong

Abstract

Precise point positioning-real-time kinematic (PPP-RTK) integrates the benefits of traditional positioning techniques RTK and PPP to provide fast and high-accuracy positioning, making it very ideal for the autonomous driving industry. Among PPP-RTK models, the Undifferenced an Uncombined (UDUC) PPP-RTK shows the capability to adeptly process multi-frequency and multi-GNSS data, demonstrating significant potential and possibilities. However, its positioning accuracy is influenced by numerous factors, particularly ionospheric delay. The PPP-RTK generates ionospheric delay corrections by the interpolation of reference stations. Therefore, it is imperative to identify an appropriate ionospheric interpolation technique to guarantee positioning accuracy. This thesis seeks to evaluate the interpolation accuracy and positioning performance of various current interpolation techniques, including Inverse Distance Weighted (IDW) method, Least Squares Collocation (LSC) method, Kriging method and Planar Fitting (PF) method, under quiet and moderate ionospheric conditions, aiming to figure out the most appropriate interpolation model for PPP-RTK in Hong Kong, which is in low-latitude regions with short inter-station distance. To achieve this objective, the author modifies the open-source MATLAB software PPPH and build up the UDUC PPP-RTK experimental platform for this thesis. Subsequently, the author resolves and analyzes the data from 16 stations within the Hong Kong reference station network. Numerical analysis shows that Universal Kriging method is the most appropriate interpolation technique for UDUC PPP-RTK in Hong Kong. This approach exhibits high interpolation accuracy and superior positioning performance in millimeter level, while also exhibiting better robustness among these methods.