AbstractFor more than 20 years, precise point positioning (PPP) has been a well-established technique for carrier phase-based navigation. Traditionally, it relies on precise orbit and clock products to achieve accuracies in the order of centimeters. With the modernization of legacy GNSS constellations and the introduction of new systems such as Galileo, a continued reduction in the signal-in-space range error (SISRE) can be observed. Supported by this fact, we analyze the feasibility and performance of PPP with broadcast ephemerides and observations of Galileo and GPS. Two different functional models for compensation of SISREs are assessed: process noise in the ambiguity states and the explicit estimation of a SISRE state for each channel. Tests performed with permanent reference stations show that the position can be estimated in kinematic conditions with an average three-dimensional (3D) root mean square (RMS) error of 29 cm for Galileo and 63 cm for GPS. Dual-constellation solutions can further improve the accuracy to 25 cm. Compared to standard algorithms without SISRE compensation, the proposed PPP approaches offer a 40% performance improvement for Galileo and 70% for GPS when working with broadcast ephemerides. An additional test with observations taken on a boat ride yielded 3D RMS accuracy of 39 cm for Galileo, 41 cm for GPS, and 27 cm for dual-constellation processing compared to a real-time kinematic reference solution. Compared to the use of process noise in the phase ambiguity estimation, the explicit estimation of SISRE states yields a slightly improved robustness and accuracy at the expense of increased algorithmic complexity. Overall, the test results demonstrate that the application of broadcast ephemerides in a PPP model is feasible with modern GNSS constellations and able to reach accuracies in the order of few decimeters when using proper SISRE compensation techniques.
Errata to “Functional models for time‐varying random objects”