Accuracy of GPS precise point positioning (PPP)

Abstract

We tried to show how the accuracy of GPS varies with respect to observing session duration by using the method of GIPSY precise point positioning (PPP), and how the site’s coordinates latitude, longitude and ellipsoidal height are affected. We used eleven IGS sites scattered almost evenly across the earth. GPS data were obtained from SOPAC archives at http://sopac.ucsd.edu/cgi-bin/dbDataBySite.cgi in Rinex (Receiver Independent Exchange) format. Processing of the GPS data was achieved through the use of Jet Propulsion Laboratory’s GIPSY OASIS II software and the application of “Precise Point Positioning” method (PPP) by Zumberge et al., (1997). In addition, least squares (LS) analysis was used to model the accuracy of PPP. Our study is based on the mathematical expression of Eckl et al., (2001) in which the dependence of accuracy on observing session is expressed with T. GPS data were segmented into shorter sessions from 1 h to 24 h. For each segment a GPS solution was produced using the PPP routine. We used RMS values from sub-segments to model the accuracy of GPS PPP. Results indicate that accuracy of GIPSY PPP depends only on the observing session T. Observing sessions shorter than 4 hour show dependency on latitude especially for the components longitude and ellipsoidal height. Using session length of 6 or more hours one can model the accuracy of GPS PPP with a simple formulation. Our results also indicate that prediction formulation for PPP is only slightly different from that of relative positioning formulation.