*Corresponding author: Li Zhang, Institute of Engineering Geodesy, University of Stuttgart, Geschwister-Scholl-Str. 24 D, 70174 Stuttgart, Germany, e-mail: li.zhang@ingeo.uni-stuttgart.de Volker Schwieger, Institute of Engineering Geodesy, University of Stuttgart, Geschwister-Scholl-Str. 24 D, 70174 Stuttgart, Germany, e-mail: volker.schwieger@ingeo.uni-stuttgart.de Research Article Li Zhang*, Volker Schwieger Improving the Quality of Low-Cost GPS Receiver Data for Monitoring Using Spatial Correlations DOI 10.1515/jag-2015-0022 received November 19, 2015; accepted February 16, 2016. Abstract: The investigations on low-cost single frequency GPS receivers at the Institute of Engineering Geodesy (IIGS) show that u-blox LEA-6T GPS receivers combined with Trimble Bullet III GPS antennas containing self- constructed L1-optimized choke rings can already obtain an accuracy in the range of millimeters which meets the requirements of geodetic precise monitoring appli- cations (see [27]). However, the quality (accuracy and reliability) of low-cost GPS receiver data, particularly in shadowing environment, should still be improved, since the multipath effects are the major error for the short baselines. For this purpose, several adjoined stations with low-cost GPS receivers and antennas were set up next to the metal wall on the roof of the IIGS building and measured statically for several days. The time series of three-dimensional coordinates of the GPS receivers were analyzed. Spatial correlations between the adjoined sta- tions, possibly caused by multipath effect, will be taken into account. The coordinates of one station can be corrected using the spatial correlations of the adjoined stations, so that the quality of the GPS measurements is improved. The developed algorithms are based on the coordi- nates and the results will be delivered in near-real-time (in about 30 minutes), so that they are suitable for structural health monitoring applications. Keywords: Low-cost GPS Receiver, Multipath Effect, Spatial Correlations Analysis, Monitoring. 1 Introduction GNSS receivers are able to measure the 3-dimensional posi- tions automatically and continuously. In order to inform the users of potential dangers as early as possible, GNSS data is evaluated block by block as a near-real-time solution. The investigations at the Institute of Engineering Geodesy (IIGS) and at other research institutions [6, 11, 14–17] show that accuracies of sub-cm can be achieved even with low-cost single frequency GPS receivers, if the carrier phase measurements of the GNSS receivers are evaluated in relative mode and the length of the baseline is up to several kilometers. Influences of baseline-length- dependent errors, such as ionospheric and tropospheric errors, can be mitigated for short baselines. The monitored objects like landslides or dams have normally an extension up to a couple of kilometers and so the low-cost single- frequency GPS receivers are suitable for these kinds of appli- cations. Apart from that, compared to the geodetic GNSS receivers which may cost more than 20,000 €, the low- cost single frequency GNSS receivers have the advantage to be cost-effective. For example, the u-blox LEA-4T / 6T GPS receivers which are investigated at IIGS cost less than 100 €. However, the influence of site-dependent errors par- ticularly the multipath effects cannot be reduced in rela- tive mode, so it is still a general problem for the precise GNSS positioning, particularly in shadowing environment. Multipath effect affects not only the accuracies of relative but also of absolute positioning and it is a limiting factor for accuracies for both geodetic dual-frequency and low- cost single frequency GNSS receivers. Since the beginning of GPS development, there has been a lot of research on the multipath effect, and dif- ferent methods have been developed to reduce the multi- path effects, such as improving the receiver technology [21], using the Signal-To-Noise Ratio [1], applying side- real filtering [3] and station calibration [23], or improving the antenna design [4, 9, 10, 18]. Up to now, there is no method which can completely eliminate the influence of the multipath effects. And many methods can only be Journal of Applied Geodesy 2016; 10(2): 119–129