Reference : GPS + Galileo Single-Frequency Relative Positioning with low-cost receivers
Scientific congresses and symposiums : Paper published in a book
Physical, chemical, mathematical & earth Sciences : Earth sciences & physical geography
GPS + Galileo Single-Frequency Relative Positioning with low-cost receivers
Deprez, Cécile mailto [Université de Liège - ULiège > Département de géographie > Unité de Géomatique - Géodésie et GNSS >]
Warnant, René mailto [Université de Liège - ULiège > Département de géographie > Unité de Géomatique - Géodésie et GNSS >]
Proceedings of the 31st International Technical Meeting of the Satellite Division of the Institute of Navigation (ION GNSS+ 2018)
31st International Technical Meeting of the Satellite Division of the Institute of Navigation (ION GNSS+ 2018)
24 octobre au 28 septembre 2018
Institute of Navigation
Miami, Florida
United States
[en] Galileo ; Positioning ; GNSS ; Single-Frequency ; RTK ; ISBs ; Low-cost ; precision ; Multi-GNSS ; Relative Positioning ; GPS
[en] During the past decade, Galileo (Europe) and BeiDou (China) satellite systems have been expanded, bringing their total number of in-orbit satellites to 26 and 35, respectively. In combination with GPS (USA), the redundancy of observations has increased, improving the precision, availability and robustness of satellite positioning worldwide.

Low-cost single-frequency receivers could largely benefit from these combinations of GNSS, as for instance in the case of safety-of-life applications, for which system failure may be critical.

In this contribution, we aimed at testing the possibility of improving the positioning performances of single-frequency low-cost receivers by using a tight combination of Galileo E1 + GPS L1 frequency. We evaluated positioning precisions of u-blox EVK M8T receivers, which cost a few hundred USDs, in multi-GNSS relative positioning based on code and phase double differences. We compared the results obtained with these low-cost receivers with those of geodetic receivers, which cost thousands of USDs.

To create a valid stochastic model for u-blox combinations, we designed a very short baseline experiment (5.6 meters) on the roof of the Physics building of the University of Liège (Belgium) in which two low-cost u-blox receivers were connected to two choke-ring Trimble TRM 59800 SCIS antennas. The use of these geodetic antennas instead of the low-cost u-blox patch antennas provides better multipath mitigation performances as well as more precise signal tracking. We had the antenna positions fixed at a few millimetre level precision so that no error coming from an a priori position estimation could influence our results. On this basis, we computed u-blox GPS L1 and Galileo E1 code and phase observation precisions by the means of double differences. First results showed code observation precisions ranging from 30 to 40 centimetres for both GPS L1 and Galileo E1 signals and phase observation precisions of 1 millimetre.

The use of a unique-pivot satellite for both constellations in GPS L1 + Galileo E1 double differences introduces differential receiver hardware delays between GPS and Galileo, also called inter-system biases (ISBs). They need to be estimated as new unknowns in an ISBs-float model. The code ISBs are estimated as an additional unknown in a least square adjustment together with the position unknowns (X, Y, Z). The method is different when it comes to carrier phase ISBs, the entire part of the ISBs being estimated all together with the integer ambiguity thanks to the LAMBDA method, so that only the fractional part (between -0.5 and 0.5 cycles) remains to be estimated as an additional error.

We first evaluated these ISBs over short baselines where low-cost u-blox receivers were connected to choke-ring Trimble antennas. In this experiment, both ISBs between u-blox receivers as well as the ISBs between u-blox and geodetic receivers were studied. The University of Liège owns 2 Septentrio PolaRx4 and 2 Trimble NetR9, which were used in this contribution. We performed an analysis of ISBs stability over time and studied the possibility of removing these errors as a constant biases in ISBs-fixed model.

We computed different values of ISBs for each different u-blox unit used in the experiment we conducted on the roof of the building. Trimble - u-blox code ISBs showed ISBs varying between 1.9 and 2.3 meters while Septentrio - u-blox code ISBS were ranging between 0.4 and 0.8 meters. Between u-blox code ISBs were also computed, showing results varying from 34 to 43 centimetres.

We quantified the positioning accuracy improvement brought by multi-GNSS single-frequency models using ISBs-fixed and ISBs-float solutions to models ignoring ISBs.

Eventually, we enlarged the baselines from a few meters to 13 kilometres. We concluded that ISBs estimated over very short baselines could be used to characterize ISBs encountered on larger baselines.

This is a first analysis of GPS L1 + Galileo E1 ISBs between low-cost receivers and between low-cost and geodetic receivers. The contribution presents reference values of observation precision and positioning accuracy over distances ranging from 5 meters to 13 kilometres. We used a software of our own, developed under MATLAB, to provide these results on 1 second data. Observations under 10° of elevation were not considered in the analysis as well as positioning solutions with bad geometry (PDOP greater than 6). We computed broadcast orbits thanks to IGS MGEX products.
Fonds de la Recherche Scientifique - F.R.S.-FNRS
Researchers ; Professionals

File(s) associated to this reference

Fulltext file(s):

Restricted access
GNSS18-0273.pdfPublisher postprint735.1 kBRequest copy

Additional material(s):

File Commentary Size Access
Restricted access
GNSS18-0273.pdf847.18 kBRequest copy

Bookmark and Share SFX Query

All documents in ORBi are protected by a user license.