Performance assessment of uncombined precise point positioning using Multi-GNSS real-time streams: Computational efficiency and RTS interruption

Abstract

The International GNSS Service (IGS) has established the Real-time Services (RTS) that support the real-time precise point positioning (RT-PPP) and related applications. Due to the increasing number of unknown parameters and network instability issues, both the computational efficiency and the RTS stream outage are the key points to realize real-time uncombined GNSS PPP. To improve the computational efficiency of uncombined PPP, the recursive Kalman based on single observation is implemented rather than the standard Kalman filter. The differences between the RTS orbits and the corresponding ultra-rapid predicted orbits reveal that the ultra-rapid predicted orbits can be used as alternatives of the GNSS real-time orbits. During the RTS clock outage period, the linear and sinusoidal terms are determined by the ultra-rapid observed part in advance, while the constant term of clock offsets is determined by the latest available RTS clock corrections. The PPP results have revealed that the proposed prediction method is much more effective than that of directly using the ultra-rapid predicted products. Even when the prediction duration reaches up to 1 h, the 3D positioning accuracy of GNSS RT-PPP is improved by 51.91% compared with the ultra-rapid scheme. Compare to the GPS-only kinematic PPP, the 3D positioning accuracy of GNSS kinematic PPP using CNES (Centre National d’Études Spatiales) RTS products is improved from 8.29 cm to 6.60 cm, and the mean convergence time of achieving 3D position accuracy better than 10 cm is shortened by about 44.99%.