No audio available for this content.
A new study compares archived real-time and post-processed precise satellite products for multi-constellation, multi-frequency precise point positioning with ambiguity resolution (PPP-AR).
Using a common processing framework, researchers evaluated single-, dual- and triple-constellation solutions with dual- and triple-frequency observations. Real-time streams achieved moderate centimeter-level positioning when corrections were continuously available, but the gap to post-processed products widened under stricter accuracy targets.
Multi-GNSS integration reduced sensitivity to product quality, whereas single-constellation PPP-AR remained strongly dependent on overall product accuracy. The results help users assess the performance of real-time precise products for high-precision positioning and understand the remaining advantages of final products in post-processed applications.
Real-time precise point positioning with ambiguity resolution (PPP-AR) is increasingly important for GNSS navigation, timing and other time-critical applications. Its performance depends on real-time satellite orbit and clock products delivered through services such as the International GNSS Service (IGS) Real-Time Service (RTS).
Earlier work shows that real-time products are less accurate than final post-processed products, especially for some constellations, while multi-frequency and multi-GNSS observations improve convergence. However, reported convergence times come from different datasets, product streams, processing strategies, and evaluation criteria, making direct comparisons difficult. Given these challenges, in-depth research is needed on how archived real-time and post-processed precise products compare under a common PPP-AR framework.
A study led by researchers at Technische Universität Berlin and the GFZ Helmholtz Centre for Geosciences, with collaborators from Chang’an University, was published (DOI: 10.1186/s43020-026-00222-y) online on Oct. 7 in Satellite Navigation. The team processed 31 days of data from 249 Multi-GNSS Experiment (MGEX) stations using archived real-time products from Centre National d’Études Spatiales (CNES) and Wuhan University (WHU), plus final WHU products. They evaluated dual- and triple-frequency PPP-AR for Global Positioning System (GPS), Galileo, and Third-generation BeiDou Navigation Satellite System (BDS-3) solutions under continuous correction availability to isolate product-stream effects.
The analysis compared real-time CNES/CLS (RT-CNES), real-time WHU (RT-WHU), and final WHU (FIN-WHU) streams. In single-constellation PPP-AR, convergence remained strongly tied to product quality. For GPS-only triple-frequency PPP-AR in the vertical component at the 10 cm threshold, the 68.3rd-percentile ensemble convergence time was about 26.5 minutes with RT-WHU and 15.5 minutes with FIN-WHU. For GPS+Galileo+BDS-3, those times fell to about 6.5 and 1.5 minutes, showing that multi-GNSS integration sharply reduces sensitivity to product-stream differences.
Yet strict thresholds exposed larger gaps. For GPS-only vertical solutions, the ensemble time was about 13.0 minutes with real-time products and 10.5 minutes with FIN-WHU at 20 cm, but rose to about 71.0 and 24.0 minutes at 5 cm. Narrow-lane (NL) ambiguity fractional-cycle residuals were far more product-dependent than extra-wide-lane (EWL) and wide-lane (WL) residuals. FIN-WHU showed the smallest orbit and clock differences, and RT-WHU generally outperformed RT-CNES, especially for BDS-3.
Ambiguity resolution improved convergence over float PPP, and triple-frequency observations helped PPP-AR more than float solutions. BDS-3 Inclined Geosynchronous Orbit (IGSO) satellites improved Asia-Pacific geometry but also increased regional dependence. At the 90th percentile and under tighter arc-wise root-mean-square (RMS) criteria, real-time streams degraded faster than FIN-WHU.
The authors said the results should be viewed as a favorable benchmark, not a guarantee of operational real-time performance. They explained that the archived real-time streams were processed offline under continuous correction availability, so correction latency, stream interruptions, and outage-induced reinitialization were not considered.. Under those controlled conditions, real-time PPP-AR can reach moderate centimeter-level accuracy, but the gap to final products grows quickly when users demand stricter accuracy. They added that multi-GNSS integration is a powerful buffer, while single-constellation users remain more exposed to product quality.
These findings matter for autonomous driving, precision agriculture, deformation monitoring, atmospheric sensing, and other real-time GNSS applications. Users with moderate centimeter-level needs may benefit from multi-GNSS, multi-frequency PPP-AR even with real-time products, provided corrections are stable. For strict accuracy targets, however, final post-processed products remain more reliable, and single-constellation users should expect longer convergence. The study also highlights the need for long-term, high-rate tests using operational real-time streams, because correction latency and outages can further degrade performance. Atmospheric augmentation and regional corrections may help accelerate real-time PPP-AR in the future for high-precision positioning, navigation and timing.
