FIG Peer Review Journal


Assessment of practical 3-D geodetic accuracy for static GPS surveying (2994)

Pasi Häkli, Hannu Koivula and Jyrki Puupponen (Finland)
Mr Pasi Häkli
Research Scientist
Finnish Geodetic Institute
P.O.Box 15
Corresponding author Mr Pasi Häkli (email: pasi.hakli[at], tel.: +358929555222)

[ abstract ] [ handouts ] [ handouts ]

Published on the web 2008-03-21
Received 2008-01-31 / Accepted 2008-03-14
This paper is one of selection of papers published for the FIG Working Week 2008 in Stockholm, Sweden and has undergone the FIG Peer Review Process.

FIG Working Week 2008
ISBN 978-87-90907-67-9 ISSN 2307-4086


We show the accuracy of static GPS using one easily readable chart. The chart covers all the conventional baseline and session lengths as well as broadcast and precise ephemeredes. This study is a part of an ongoing project that studies the quality of geodetic GPS at the Finnish Geodetic Institute. We created a test field that covers distances between 0.6 and 1069 km. All the test points were either permanent GPS stations or concrete pillars. This ensures that the results are free from centering and height reading errors. The ITRF2000 coordinates for the test field were processed with the Bernese software using at least 24 hours of data. These coordinates were used as reference coordinates in the test. The set of data used in the study is a random sample chosen from the data from several GPS campaigns. This way it was to give a realistic picture about the variations e.g. in annual and daily periods of GPS. Session lengths vary between 10 min and 24 hours covering the distances up to over 1000 km. The data was processed with commercial GPS software using standard processing parameters. The accuracy is presented for individual baselines i.e. adjustments were not applied. The results were pre-processed in order to eliminate gross errors from the surface fit. A 3-D grid was formed from the data so that each grid point represents one baseline and session length. Each grid point is a rms-value that contains 20 independent baseline solutions with different data. A 3-D surface was fitted over the rms values. Since the data was rather heterogenious a series of fitting schemes were tested and the one with the best fit was chosen. The R^2 value for the best fit was 0.91 for broadcast and 0.87 for precise orbits. As a result we generated a chart with logaritmic session length and baseline length axes. The rms –accuracies are shown in the chart as 1-5 cm contour lines for both broadcast and precise orbits.
Keywords: GNSS/GPS; Positioning; accuracy