High-sensitivity (HS) GPS receivers offer potentially unprecedented availability and under certain circumstances even better accuracy than conventional receivers. They are less sensitive to signal attenuation, can track weaker signals and may provide more observations. However, signal attenuation is usually associated with signal distortion and corresponding larger errors. While typical pseudo-range errors of HS receivers are on the 4– 154 ION GNSS 18th International Technical Meeting of the Satellite Division, 13-16 September 2005, Long Beach, CA 10 m level, we have actually also found errors of several tens to hundreds of meters. Any potential gain in availability and accuracy may therefore be annihilated by the effect of uncorrected errors, unless the different variances of the observations are taken into account properly. Failure to do so may result in systematic biases of several hundred meters in the position domain and subsequent errors like wrong road segment identification in vehicle navigation or erroneous geo-referencing in GIS data collection. We show that a variance model based on the signal-tonoise ratio i.e., the SIGMA-e model, accounts for the different variances. This model was originally developed for GPS carrier-phase observations but performs exceptionally well also with HS pseudo-range observations, as we show here. SIGMA-e can correctly model more than 99% of the L1 pseudo-range errors in benign environment and still about 80–90% in a worst case environment. SIGMA-e generates little computational load and is very well suited for kinematic and real-time applications because it is a simple function of the measured C/N0. It outperforms equal or elevation dependent variances by far and should be the model of choice when processing HS GPS data. Using real data obtained from three different HS GPS receivers, we show that the use of the SIGMA-e variance model improves the accuracy of the estimated positions by typically 30% to 50%, and sometimes even more.
[1]
Andreas Wieser,et al.
Robust and Fuzzy Techniques for Parameter Estimation and Quality Assessment in GPS
,
2002
.
[2]
Vijayan N. Nair,et al.
7 - Methods for Assessing Distributional Assumptions in One- and Two-Sample Problems
,
1994
.
[3]
Mark G. Petovello,et al.
Failure Scenarios to be Considered with Kinematic High Precision Relative GNSS Positioning
,
2004
.
[4]
Jinling Wang,et al.
Stochastic Modeling for Static GPS Baseline Data Processing
,
1998
.
[5]
Clyde C. Goad,et al.
On optimal filtering of GPS dual frequency observations without using orbit information
,
1991
.
[6]
B. Ripley,et al.
Robust Statistics
,
2018,
Wiley Series in Probability and Statistics.
[7]
Per Enge,et al.
Fundamentals Of Signal Tracking Theory
,
1996
.
[8]
Christian Tiberius,et al.
VARIANCE COMPONENT ESTIMATION AND PRECISE GPS POSITIONING: CASE STUDY
,
2003
.
[9]
H. Hartinger,et al.
Variances of GPS Phase Observations: The SIGMA-ɛ Model
,
1999,
GPS Solutions.
[10]
Lawrence Lau,et al.
Improvement of GPS Relative Positioning Accuracy by Using SNR
,
1999
.
[11]
R. B. Langley,et al.
P OSSIBLE WEIGHTING SCHEMES FOR GPS CARRIER PHASE OBSERVATIONS IN THE PRESENCE OF MULTIPATH
,
1999
.
[12]
Kai Borre,et al.
Time Series Analysis of GPS Observables
,
2000
.
[13]
Peter Teunissen,et al.
Quality Control and GPS
,
1998
.
[14]
Peter Teunissen,et al.
A recursive slippage test for use in state-space
,
1989
.