Raspberry Pi for GNSS-IR; A review on a recently-published paper

A new study has been recently published, in which a low-cost GPS antenna connected to a Raspberry Pi system has been employed to conduct a GPS-IR experiment for water level monitoring. In this paper, the authors have described the advantages of this system; however, there are some hidden points that I am going to flesh them out.

Makan Karegar, who is currently a researcher with the University of Bonn, is known for his innovative studies in various areas of GPS applications, ranged from gravimetric modelling to COVID monitroing. In his new paper, Makan, along with three other well-known reflectometrists (do we ever have such a word? reflectometrist?) analyzed the potential of low-cost GPS antennas attached to a Raspberry Pi microcomputer for water level monitoring. This system, which is called RPR (Raspberry Pi Reflectometer), was installed next to a river in Germany co-located with a guage for an interferometric reflectometry experience (you may watch this animiation of mine to see how GPS-IR works for the height estimation), and showed an acceptable RMSE (1.5 to 3 cm). Makan also properly mentioned how the advanteges of this system outweigh its limitations in both technical and financial ways. However, I would like to add three points based on experiences that I had in the past in working with a similar RPR. If you’ve ever been the reader of this blog, you probabely remember some of my experiences with RPR for soil moisture monitoring. I also had an unsucessful experience of using RPR for lake ice thickness and extent monitoring, which I will elaborate why it failed. So, I thought that I may be eligible to put some comments on this newly-published and, of course, valuable paper.

Firstly, the authors have emphasized the affordable price of this system compared to other GPS reflectometry sensors. That’s correct. Compared to what RPR gives, its cost (~$200) is reasonable; but it’s not THAT cheaper. The authors have mentioned that a very high-quality geodetic GPS receiver is higher than $10,000, which is correct, but one doesn’t need such an instrument for GPS-IR purposes. As an example, in 2019, I conducted a GPS-IR experiment using a brand-new EMLIS RS+ GNSS unit, which cost ~$700 (tax and shipping included). Besides, we didn’t have to develop any code to extract the navigation data since the unit simply gave us RINEX files, which could be directly used as an input for the gnssrefl software. More details about our experiment can be found in my M.Sc. thesis.

Secondly, the whole RPR system, as tabulated by the authors, is indeed low-cost. Therefore, one can connect an extra antenna with the left-hand polarization at a very low price too. The details of such a dual-polarized instrumentation has been provdied in Gary Chan‘s M.Sc. thesis, and we use the same system for our soil mositure experiment, as well as that failed lake ice research. The benefit of this designation, compared to the only-one-antenna one, is obvious for reflectometrists; no need to discriminate the direct and reflected signals, chance to receive reflections from satellites with higher elevation angles, and opportunity for comparing GPS-IR and GPS-R.

Thirdly, RPR is fragile, and that’s why our lake ice experiment failed: the RPR may work in normal temperture, but it can breaks in very cold conditions. To be more specific and clear, we had planned to install our RPR to a tower located on the shore of a lake in central Ontario to collect reflections from lake ice surface. Although, we’d tested our system in a lab fridge at -80 C to make sure it would work at very cold conditions, its USB connections broke after installation. So, before conducting any experiment, the tempeature variation of the river, or any other study site, should be taken into consideration in order to plan for any possible freezing sitatuons. The authors have wisely employed a heat sink pack to cool the system in the summer; so, it would be reasonable to think about an opposite component for the witertime as well.

Let’s look forward to seeing a network of RPRs soon, which, as suggested by the authors, is indeed technically and financially doable.

GIRAS; a GNSS-IR Analysis Software

Recently, a GNSS-IR analysis software has been published online offering a handy graphical user interface, which tends to become a popular alternative to existing software packages and tools. This MATLAB-based open-source software package is GIRAS.

GNSS-IR has become a robust method to extract the characteristic environmental features of reflected surfaces, in which, fluctuations in the strength of the signal received by the GNSS receiver are analyzed. A short video was created and posted here showing how fluctuations in reflected SNR relates to altimetric applications. Frequency, amplitude, and phase values that provide the most appropriate model of the changes in signal strength are used as GNSS-IR metrics. Although there are several different software packages currently available to find these metrics, they may be insufficient in some cases.

Newly, a MATLAB-based, open-source GNSS-IR analysis software (GIRAS), developed by Cemali Altuntas and Nursu Tunalioglu from Yildiz Technical University, Turkey, has been published online in GPS Solutions. With its graphical user interface and capabilities, GIRAS is a good alternative to existing software packages and tools. GIRAS has three main modules and five sub-modules. The first main module is “Read & convert files”. This module reads raw GNSS data and stores the necessary observations in the MATLAB environment. RINEX version 2 and RINEX version 3 observation files are supported. Both broadcast ephemeris and precise ephemeris (sp3) files can be used. Multi-GNSS (GPS, GLONASS, Galileo, Beidou) sp3 files are supported.

The second main module is “Pre-analysis” which has three submodules: (1) SNR & dSNR data, (2) Sky view, (3) FFZ. Here, the user can plot the SNR and dSNR data for different polynomial degrees for each satellite and signal type. The sky view can be plotted as including the selected satellite systems. The first Fresnel zones (FFZs) can be displayed on the plot screen and Google Earth. The user also can export all graphs as MATLAB figures.

The third main module is “Analysis”, and it has two submodules: (1) Make estimations, (2) Improve estimations. Here, GNSS-IR metrics can be estimated using several options included. In addition, filtering and outlier analysis of the results can be performed. Analysis results can be saved in both TXT and MAT formats.

GIRAS enables the selection of azimuthal and elevation angle masks with multiple range selections for either short-term or long-term GNSS data collected, which may help to predict future models for climatological studies. The software is also suitable for multi-GNSS analysis. The quality-control modules are in a way of multiple choice, in where the researchers can apply options for their aims of use. The output files are well-designed for users in a defined format for further computations.

GIRAS seems to be widely used in future GNSS-IR work, thanks to its functional GUI and many options included.