Study on GPS–PPP precision for short observation sessions

Precise point positioning is increasingly being used in geodetic applications that in many cases are based on static 24-hour RINEX files. Since there are many applications where sub-centimeter position accuracy is not required and users wish to use a single receiver and not be dependent on differential correction, we will evaluate PPP performance for static positioning with 12-, 6-, 3-, 1- and ½-h observations. We have, therefore, considered a dataset for the year 2013 from 14 European GNSS stations. The data were analyzed using GIPSY-OASIS II software package and evaluated in terms of repeatability of the coordinates and of coherence with the formal error indicated for each PPP solution. Particular attention was paid to solutions showing large discrepancies in coordinates. The test shows that PPP precision for the 24-h files is below 5 mm, but decreases slightly for the 12-, 6- and 3-h observation sets. For the 1-h and the ½-h RINEX files, precision is within 5 and 10 cm, respectively. The analysis is completed with a discussion on the impact of the ambiguity resolution that shows how it significantly improves only the easting component and moreover has a higher influence on the formal error rather than on the solutions. Lastly, the study contains an investigation into the reliability of the formal error associated with the PPP solutions. We show that the formal error can be used to identify incorrect solutions, but is not suitable to represent the real accuracy. For that reason, we propose to use the formal error given for the float solutions even for the ones with fixed ambiguities.