When computing post-processed PPP solutions, I typically had to pick one of the two following options: a GPS-only solution with ambiguity resolution (only available from CNES) or a GPS + GLONASS solution with float ambiguities (available from several other analysis centers). This dilemma ended last week, with CNES also offering satellite clock corrections at 30-sec intervals for GLONASS and Galileo. This product is provided as a part of the MGEX project, but the GPS/GLONASS part will soon become the standard GRG product.
To test the impact of adding GLONASS to a GPS PPP-AR solution, I processed one-hour segments of a 24-hour data set from station UNB3 in Fredericton, Canada, collected on 15 March 2015. Each session was processed in kinematic mode, i.e. making no assumption on receiver dynamics. The figure below shows that the GPS-only solution (in black) converges to cm-level accuracies in typically less than 30 minutes, except for one session requiring over 45 minutes. Adding GLONASS (in red) reduced convergence times to about 10 minutes, with only one session exceeding 20 minutes. Further testing is obviously needed to draw significant conclusions, but those results are promising.
With Galileo launching new satellites and a third frequency on the way, we can simply wait and watch fast PPP convergence become a reality.