Session B3: Precise GNSS Positioning Applications

Improving the Performance of Low-Cost PPP Using Fugro’s NextG4 Service
Dennis Odijk, Fugro, Australia; Xianglin Liu, Yahya Memarzadeh, and Artur Oruba, Fugro Innovation & Technology
Location: Beacon B

Fugro’s long-standing history of providing global GNSS positioning capability has evolved into its NextG4 service (also known as G4+), in which real-time corrections are transmitted via L-band from geostationary satellites to enable Precise Point Positioning (PPP) based on GPS, GLONASS, Galileo and BeiDou. Next to precise orbit and clock corrections the service provides carrier-phase bias (also referred as Uncalibrated Phase Delay) and code bias (also referred to as Inter-Frequency Code Bias) corrections at three frequencies per constellation such that high-accuracy (centimeter-level) PPP based on integer ambiguity resolution (IAR) becomes possible (the so-called PPP-IAR solution). So far Fugro’s PPP products are applied to correct high-grade data of many offshore users who traditionally operate geodetic receivers. With NextG4 a 95% positioning accuracy of 2.5 centimeter horizontally and 5.0 cm vertically is achieved within a convergence time of three minutes for such high-grade receivers. This convergence time is hampered mainly by ionospheric delays which are not corrected.
In this contribution we demonstrate for the first time the performance of Fugro’s NextG4 corrections when applied to low-cost GNSS devices operating in challenging environments. Challenging in this context means a dynamic receiver that is attached to a car moving in a suburban environment such that signal blockage and reflections frequently occur, thereby causing cycle-slips and multipath effects. We have used data from the u-blox ZED-F9P module, which is a multi-frequency mass-market receiver equipped with a low-cost patch antenna. The performance of this low-cost receiver is measured in terms of positioning accuracy in combination with the convergence time needed to reach the desired accuracy. The desired positioning accuracy together with the convergence time one is willing to accept drives the type of corrections that is needed.
As benchmark for our analysis, we first show the performance when the “best” corrections are applied, i.e. NextG4 orbit and clock corrections together with PPP-RTK corrections for phase and code bias, as well as ionospheric and tropospheric delays that are determined from a nearby (< 1 kilometer) high-grade reference receiver. The results demonstrate that using the PPP-RTK corrections from the nearby station horizontal centimeter-accuracy is achievable for the u-blox within one minute. Without the phase and code bias corrections a PPP-IAR or PPP-RTK solution cannot be computed, but still a standard PPP solution is possible, having an accuracy at the few decimeters level, which is however quicker available than a PPP-IAR solution, i.e. within few seconds, thanks to the high-quality nearby ionospheric corrections.
In a next step we compare this benchmark solution with a true PPP solution, thus without data from a nearby reference station. For this we use phase and code bias corrections from Fugro’s NextG4 service, together with the orbit and clock corrections. As the NextG4 service does not provide ionospheric corrections, but these are essential for the convergence time, we compute real-time Global Ionospheric Map (GIM) corrections based on GNSS data from over a hundred Fugro’s reference sites having a spacing of hundreds of kilometers. In addition, corrections for Differential Code Biases (DCBs) are computed from these reference stations. Since these GIM+DCB corrections are much less accurate than the nearby ionospheric corrections the PPP results are worse than the benchmark results, something which could already be expected: the horizontal positioning accuracy reduces to submeter level, while the convergence time to reach this accuracy can be up to few minutes. We emphasize that the inclusion of the NextG4 phase and code bias corrections does not help here, as without precise ionospheric corrections rapid integer ambiguity resolution is not possible and a PPP-IAR solution cannot be obtained.
If one does not accept a horizontal positioning accuracy at submeter level that is reached after a few minutes, the solution lies in the use of more precise ionospheric corrections, such that by making use of Fugro’s NextG4 phase and code bias corrections a PPP-RTK solution becomes feasible for the u-blox. Although it is realized that a denser network of reference stations than used to generate the GIM corrections may better capture the spatial variations of the ionospheric delays, for this contribution we analyze the suitability of ionospheric corrections determined from a single reference station as to investigate what is still achievable in terms of PPP for low-cost receivers. For this we have used a reference station within ten kilometers from the u-blox receiver as well a much further (tens of kilometers) reference station. We investigate the effect of these different sets of ionospheric corrections on the PPP-RTK performance of the moving u-blox. As a reference of what is feasible in terms of PPP integer ambiguity resolution, we compare the performance of the low-cost receiver to that of a high-grade receiver and antenna attached to the same car.