Post-procesing software for smartphone RINEX files

V. Kelly Bellis

unread,

Apr 20, 2021, 11:48:11 AM4/20/21

to GPSTest

Could anybody please provide any information on software you're using to post-process RINEX files that are being produced from your mobile device(s)?

From what I've tried using GnssLogger's produced RINEX files, I've been unable to find any online service or stand alone program that permits post-processing.

Thanks for any reply.

Kind regards,

Kelly

Sean Barbeau

unread,

Apr 20, 2021, 11:54:05 AM4/20/21

to V. Kelly Bellis, GPSTest

FYI, I maintain a list of services/libraries I know about here:

https://github.com/barbeau/awesome-gnss

I believe several of them support RINEX, although I haven't tried them myself.

Sean

--
You received this message because you are subscribed to the Google Groups "GPSTest" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gpstest_andro...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gpstest_android/9f42f06c-ffc6-479e-a2c6-9d16928c3150n%40googlegroups.com.

Валерий Заподовников

unread,

Apr 20, 2021, 11:54:12 AM4/20/21

to GPSTest

Rtklib, of course. And gnss-sdr too! https://github.com/tomojitakasu/RTKLIB/tree/rtklib_2.4.3

вторник, 20 апреля 2021 г. в 18:48:11 UTC+3, kelly...@gwi.net:

gpsfan

unread,

Apr 20, 2021, 12:16:24 PM4/20/21

to GPSTest

Yes RTKLib will work but the results are going to be poor (don't expect to get an RTK Fix) due to the poor antenna designs. RTKlib expert "rtklibexplorer" took a crack at PPing experimental data collected using a choke-ring antenna here https://rtklibexplorer.wordpress.com/2021/02/27/another-look-at-l1-l5-cellphone-ppk-with-rtklib/

Jason is an online also PP Android raw data https://jason.rokubun.cat/ but they don't show the quality of the solutions (fix/float/single) even though they use RTKLib as the processing engine so you don't really know what you're getting.

If you want to experiment with post-processing you're better off buying a $15 uBlox 8 USB receiver on AliExpress !

V. Kelly Bellis

unread,

Apr 20, 2021, 1:19:25 PM4/20/21

to GPSTest

Since retiring and selling my surveying business, I no longer own any survey equipment, but can still improvise :). The wooden base shown in the photo above have since been wrapped with aluminum foil to help mitigate multipath.

@gpsfan - My current static tests are using (L1 only) Samsung A71 on a surveyed control point and GnssLogger v3.0.0.1 but at this point, I'm just trying to see if the RINEX files that GnssLogger are producing can even be post-processed. I've always involved NRCAN's CRSR-PPP in all of my experiments ever since I first learned their service 20 years ago; L1 only (Ashtech) and L1/L2 (Javad). And a number of these last smartphone RINEX tests produced by Google's GNSSLogger (shared Drive folder here, in case you want a peak) were submitted to NRCAN's CRSR-PPP service, but none of them were processed and the results returned to me; not even a notice of failure. That led me first to trying to check on the validity of the RINEX files' formatting. Javad has a very old tool, Rinex Files Viewer (beta version 1.0.6), which is no longer supported, but still seems to work. It showed errors in the sample RINEX file, but not in any fashion I care to share with the Google developers. That was a good thing as it next led me to GFZRNX (Windows 64-bit cmd line program that's wholly contained in gfzrnx_1.15_-8044_win64.exe) (gfzrnx_1.15-8044_manual.pdf), a wonderful tool box that not only can report errors, but can fix them. The bummer was the repaired RINEX file sent to CRSR-PPP never returned any results either. I've written to the kind folks at NRCAN asking if smartphone RINEX files were permissible. I'm guessing that it's the poor smartphone antenna that will preclude such submissions from being processed, but for now, that's only a hunch. And while I await their response I tried using my ancient GNSS Solutions from back in the Ashtech days, but it needs the observation's (2) accompanying PM2 files (met and nav). Thus my bugging this group for a more contemporary solution.

@val - Thank you for the link. Might you be willing to see if the attached log will even work with RtkLib?

@Sean - Thank you for maintaining the list at https://github.com/barbeau/awesome-gnss

Kind regards,

Kelly

gnss_log_2021_04_18_12_08_09.zip

Sean Barbeau

unread,

Apr 21, 2021, 1:07:40 PM4/21/21

to V. Kelly Bellis, GPSTest

FYI, note that Google has posted some of their own raw GNSS measurement datasets here:

https://www.kaggle.com/google/android-smartphones-high-accuracy-datasets

...for the ION GNSS+ contest:

https://developer.android.com/guide/topics/sensors/Google_Open_Dataset_Challenge.pdf

It includes RINEX data.

Sean

--

You received this message because you are subscribed to the Google Groups "GPSTest" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gpstest_andro...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/gpstest_android/fc9d383c-5fbd-4caf-974c-c6e2e8e2218dn%40googlegroups.com.

V. Kelly Bellis

unread,

Apr 21, 2021, 2:39:21 PM4/21/21

to GPSTest

@Sean - Thanks for the heads up - looks interesting.

gpsfan

unread,

Apr 22, 2021, 3:54:25 AM4/22/21

to GPSTest

Yes I'm not sure why your files are not being processed, maybe because dual frequency is required for PPP ?

PPK will work for sure with L1 only, to start with you can you upload your file on Jason https://jason.rokubun.cat/#!/upload (designed to be Android GNSSLogger compatible) and see what happens. Then if you have some time you can look into RTKLib using the tutorials/settings provided on the RTKLibexplorer blog, he has a recent article specifically about Android L1/L5 smartphones.

V. Kelly Bellis

unread,

Apr 22, 2021, 3:46:40 PM4/22/21

to GPSTest

@gpsfan - Thank you very much for the Rokubun link. That service is quite interesting! I've written to Isaac Hoyas Ester with some questions and comments and I await his reply. In other softwares; e.g., Topcon Tools, Ashtech Solutions, GNSS Solutions, and PPP services from https://webapp.geod.nrcan.gc.ca/geod/tools-outils/ppp.php, I have always successfully post-processed single frequency static observations (made with survey-grade equipment; e.g., Ashtech ProMark 2s) against multiple CORS, and with long observation times, these positions have typically been excellent, within a centimeter. The PPP solutions from NRCAN also produced excellent results from L1 only observations made with survey gear, but not from the observations from my Samsung A71 using GnsssLogger, and NRCAN never provides any notice as to why any of these smartphone RINEX file submissions fail to process a solution. A couple of the questions I asked Isaac: Why can't Rokubun use CORS data for single frequency smartphone ? and Why can't Rokubun perform network solutions; i.e., using multiple baselines (multiple CORS)? I dunno, but still, I'm impressed with Rokubun. See test example below. I also asked him if Rokubun is able to accommodate user's Height of Instrument (HI) entry prior to processing? I'm trying to lobby GnssLogger's developers to allow entry of the HI in their app at the time of observation, but I haven't gotten any response. Note that in the (below) comparison of Rokubun test results with surveyed control data (OPUS solution available upon request), adjusted for HTDP and in context of IGS08, 1.18 meters should be subtracted from Rokubun's listed Ellipsoid height; i.e., 37.48 - 1.18 = 36.30 (smartphone) vs 19.32 (surveyed) ellipsoid height.

Yes, Val also mentioned RTKLib, and I have been looking at RTKLib, but I must confess, I've yet to start playing around with it as its presentation is a bit perplexing to me. For example, the most recent release notes are from 8 years ago http://www.rtklib.com/rtklib.htm. Reading the readme.txt https://github.com/tomojitakasu/RTKLIB_bin/blob/master/readme.txt for the most recent stable version, or https://github.com/tomojitakasu/RTKLIB_bin/blob/rtklib_2.4.3/readme.txt for the most recent beta, it's quite terse: "The binary APs and DLLs for Windows." I asked Val if he'd try processing the GnssLogger observation file just to see if it would work since he seemed already familiar with using RTKLib. From looking at the manual from 2011, it looks like it was intended for running on Windows XP. I might write to Mr. Takasu next to make sure I'm on the right track.