Whats the best way to log raw RTK data? --- piksi_sample_grabber, peregrine,pyNEX,sbp_example

[Marvin Grey]

Hello everyone,

as many of you I am also trying to get hold of DGPS/RTK positioning data but as it seems piksi is not quite there yet. Piksi is a developing positioning tool, but still many of the customers are hoping that they can start precise surveying soon. To be able to use Piksi I would like is to get a file where all necessary RTK data for later post-processing is saved. I would not assume that I would be able to understand such a file at the moment, but I want to be sure to have all the information logged, so I can process them to actual coordinates whenever the project and myself is ready for this.  So what would be the best and easiest way to do this ? (the log-files created by the console.py are only about 10m accurate and therefore not appropriate)

- in the FAQ it says:

How can I log coordinates?

How to use the Swift Navigation console, please read Piksi_User_Getting_Started_Guide. Point logging functionality will be added shortly. In addition, you can use any program that accepts NMEA or RTCM.

My Question: when is point logging function ready ? And will this be the best positioning data possible?

-also in the FAQQ:

Can you record data from two devices and post-process later into RTK solutions?

Yes. Observations can be converted into RINEX for post processing with RTKLIB or another RTK software.

My Question: How can I do this ? And what has pyNEX to do with it?

- the sbp_example code provided here: https://gist.github.com/fnoble/8b1b12a1d15b26c5546a has gotten me quite far, but I am having trouble printing out all information (BaslineNED, BaslineECEF, PosLLH, PosECEF ) at the same time, also all 'acuraccy' fields are printed as 0.000. Would it be possible to provide a similar python function which saves ALL necessary data into a log file ?


-the most elaborate possibility would be 'peregrine', as far as I understood. But I cant get libswiftnav-python installed. And even if I could what can I expect from peregrine?

-eventually there is already a solution for my problem, but before I get into the installing procedure : What does the piksi_sample_grabber do ?

Please post your knowledge/experience so we can get closer to our applications.

Greets
Marvin

[Clive Turvey]

Well for post-processing you'd be looking to generate RINEX .obs and .nav files for the Base and Rover, I'm not sure the RTK solution is particularly helpful for this as it's an "on-the-fly" computation, and post-processing is a "with hindsight" computation which looks at the satellite measurements, not the navigation solution generated from them. The console permits you to record the observations, you'd then need to select some post-processing tools which will accept the data. If you've never done post-processing, that would perhaps be a good place to start your studying and experimentation.

Are you looking for others to be coding the python for you, or direction?

[Marvin Grey]

Thanks, I think that is the right direction. So I managed to install RTKLIB , Ive got it on Ubuntu and on my Windows 7 VM because there the GUIs work. With RTKPOST 2.4.2 (similar to RNX2RTKP on linux), trying all positioning modes, I used the obs. RINEX data generated by the piksi console for both ROVER and BASE, but I always got : nav error.RTKPOST asks for nav. data ,just as Clive said , but where to get it ? Piksi gives us eph. data, which can be used for certain positioning modes in rtkpost but it still wants nav files.

As it says in the RTKLIB Support for an earlier version :

No.34 A RINEX GPS NAV file is always needed for post-processing (RTKPOST, RNX2RTKP ver.2.3.0)

At least one RINEX GPS NAV file is always needed for post-processing even if the options only use GLONASS or use precise ephemerides. If no RINEX GPS NAV file is specified, the post-processing always is terminated as no navigation data error.

Due to the bug in src/postpos.c. (2010/01/28)

Fixed for GLONASS only positioning in ver.2.4.0. With precise ephemerides, broadcast ephemerdes are still needed to get rough clock parameters, ionosphere parameters and TGD.

Does Piksi really deliver 'precise ephemerides' ? What are these and what are nav files ?

Thanks in advance for all contributions!

Marvin

[Marvin Grey]

[Clive Turvey]

Depends where you're situated. In the US you can get NAV (broadcast ephemeris) data from the NGS/NOAA/CORS site, which is an amalgam of broadcast ephemeris (from the satellites) received at all CORS locations. ie ALL rather than just the ones visible to you.

Precise ephemeris is data provided in .SP3 files and provides exact positional information after the fact, rather the the curve fitted forward projected estimates broadcast by the satellite, and periodically uploaded the constellation. These in current times are uploaded regularly, but the constellation can work autonomously for a significant time without them. ie war or force-majeure. These SP3 files are also published along with the daily RINEX data on the NOAA ftp site.

ftp://www.ngs.noaa.gov/cors/rinex/2015/012/

[Marvin Grey]

Okay, so  I got the nav data with RTKGET from IGS_NAV repository. Actually I couldnt get any good information which broadcast ephemeris repo I should take for a receiver placed in Berlin, but BKG wouldn't work - so I took the one attached. RTKPOST actually worked ! it produces position files like the one attached : kin_bro_cont_thr3_trioncor  -- the name reflects the various options I tested. But the results were poor. Just as the position_log files produced by PIKSI they are about 12m accurate :-( ... The eph file from Piksi - what is it good for - it was not used by RTKLIB!

How can I get what swiftnav has promised ? Can anyone show me accurate positions with the files I provide here ? I am giving up my privacy - if you get the position you can send a drone on top of my flat! But please tell me how you managed to get accurate positions...

Marvin

[Clive Turvey]

Static post processing base against rover, and rover against base

Base Location  : 52.522762096   13.334861359    95.0855

Rover Location : 52.522769129   13.334864056    96.7729

This was done quickly with RTKPOST, my preferred software doesn't like the locked pseudo-range the Piksi outputs.

For CORS data you'd look to http://www.sapos-bw.de/

[Marvin Grey]

Wow! Thats what I wanted 0.839 m distance thats similar to what was on the piksi console. On my roof the distance was 0.60 cm. But that is still okay. Thanks, if you are around visit me for a beer! I understand much more now, and I have started hacking the console code to get hold of the information I want. I will write the essence of my survey plan and if you still have time, I would appreciate your help and comments:

When I go in the field (archaeological survey in Saudi-Arabia) I will get good CORS data for that region.
I will setup the piksis, one fixed on top of the car, the other as mobile tool,
I wait for rtk fixed, than  log at least the Baseline NED values and record the observations of both piksis.
Than I use rtkpost to post-process the position of the piksi on the car against the mobile position of the tool-piksi. I use 'base move mode' or another mode? Do I set 'precise' or just 'broadcast'.
The .eph files that piksi brings us downloaded from the sattelites are not used !?
Than I can translate the car-position to any projected coordinate system, and I will get the mobile tool position by adding the NED-values of the baseline to the car-position.

What do you think about the Baseline output of the piksis? probably I could also use only post-processing... What about the python BaselineECEF output? could I use only this to retrieve a accurate position for both piksis?