[FOSS-GPS] first steps with RTK and RTKPOST

henri Debise henri_debise at yahoo.fr
Thu Nov 26 10:31:14 PST 2015


I'm new to the foss-gps list. 
I come from France, so please excuse me if my English is not very good. 

I work in a small company which works with drone, the question of georeferencement is in our core business. 
I'm currently studying the best choice to increase precision of gps track log. 
So I'm studying results from PPP free service from canadian natural resources : http://webapp.geod.nrcan.gc.ca/geod/tools-outils/ppp.php?locale=enI'm also studying RTKLIB. My questions are mainly related to RTKLIB, but as GPS technology is quite new for me, I may misunderstood basic questions.

On my test survey, we are using RTK mode:  a base station in installed close to the drone (300 meters distance), and broadcasts live corrections to the drone which also has a GNSS antenna. The base is not installed on a known point, and we don't let the base a long time to increase the accurate of the base position. We use the base for 20 minutes only. 
As the position of the base is not accurate, it is necessary to post process the GPS track log of the drone, then with RTKPOST

As I'm in France, I've downloaded Rinex correction file from permanent GNSS station (http://rgp.ign.fr/DONNEES/diffusion/). I've downloaded a file with a frequency of 1 second. I got observable and navigation file. Then I use RTKPOST from RTKLIB. 
In field 'RINEX OBS : rover', I input my Rinex observable rover file (I don't use rtknavi).  
in field 'RINEX OBS : base station', I input the observable file from the permanent GNSS base station (rgp.ign). 
in field 'RINEX *NAV...' I input the navigation file from the permanent GNSS base station (rgp.ign). 
In options, Setting1 , I define the positioning mode to Kinematic. 
Frequencies are L1+2, forward. Elevation mask is 10°. 
in field Ionosphere correction, I tried 2 different inputs (without significant differences on the result) : Broadcast, and Iono-Free LC. 
Troposphere correction is set to off and Satellite ephemeris/clock is set to broadcast. 
I only work with GPS (5 or 6 satellites are available during the flight). 
I fill the coordinate of the permanent GNSS base station (rgp.ign) in Positions tab. I let all the other input to default. 
Then I execute. 
different things about the results : 
- the quality flag, Q, is set to 5: Single, solution by single point positioning. 
Shouldn't it be equal to 1 or 2 ? 
- I wonder why, in the input of the rinex observable file, we can't select many files ? (same question for rinex navigation file)I mean, I can have many files, each one from a different permanent GNSS base station. Do I have to merge those files before processing in RTKPOST ? 
- In my results, I get a gap of around 5 meters, in comparison to my RTK gps track log without post processing. This looks to me as a very important gap. I can't say if the result is better, but is it possible to get a so important gap between RTK with and without postprocessing? In my rinex observable file from the rover , I had 1 and 1 for WAVELENGTH FACT L1/2, that's enough to ensure RTK mode had really worked ? 
- Also I have noise, on the post processed track log. Small noise : around 10 cm for 0.3 second for example, but this noise does'nt exist on my RTK gps track log without post processing. And the change of ionosphere correction to Iono-Free LC didn't successed to remove the noise. How can I remove the noise ? 

I heard that PPP solution is not suitable for my configuration as the solution as no time to convert, the flight is too short. If you have any comments about this, there are welcome. 

Thank you for your attention, I hope to read soon any comments/suggestions/.... 


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/foss-gps/attachments/20151126/84d0b7bc/attachment-0001.html>

More information about the FOSS-GPS mailing list