Uncombined precise point positioning with triple-frequency GNSS signals

Abstract

With the modernisation of GPS and the development of GALILEO/BDS, more satellites can transmit multi-frequency signals, which brings new opportunities and challenges for data integration in multiple systems. This work focuses on the performance of triple-frequency precise point positioning (PPP) and the phase anomaly phenomenon due to the introduction of the third frequency. First, three PPP models, particularly the triple-frequency PPP model using uncombined observations, are derived. The new biases, inter-frequency biases (IFBs), are estimated to compensate for the pseudorange hardware delays in the triple-frequency PPP model. Then the reason for the phase anomaly on the third frequency is analysed theoretically. Finally, three PPP models with real GPS/GALILEO/BDS triple-frequency data are performed to evaluate the performance of triple-frequency PPP and the influence of time-variant phase hardware delays. Due to the smaller magnitude of the time-dependent phase hardware delays for GALILEO and BDS, even regardless of them, the triple-frequency PPP model can achieve a similar or better positioning accuracy. However, this is not the case for GPS. The results show that the traditional clock products based on dual-frequency ionosphere-free (IF) combinations cannot be directly used in the GPS triple-frequency PPP because of the time-dependent part of the phase hardware delays. Ignoring the time-dependent part will cause a much poorer positioning performance compared with dual-frequency PPP. After adding a small amount of process noise to the GPS ambiguities on the third frequency, the 3D positioning accuracy of the GPS/GALILEO/BDS triple PPP model can achieve marginal improvement for both static and kinematic mode.