Struggling with PPK Workflow

Hi all,

I’m attempting to perform a PPK workflow using 2 RS2 units and a CORS network, but I’m struggling to determine the location of the base. When attempting to process the base observation against the CORS observation in RTKPOST I only get a 20-30% fix that varies 5-10cm. I’ve tried adjusting every setting I could find but I haven’t been able to achieve a usable result.

Would somebody with more experience be able to take a look at the data and see if they can identify what’s wrong? When reviewing the base and rover logs everything I see appears to be fine (numSats, SNR, etc). The Base, Rover, and CORS logs can be found here: Onedrive Link

I’m using RTKLib ver.demo5 b31 with the latest version of reachview on the RS2(not sure of the exact version, they’re still out of the country). I was able to successfully test this workflow last week but now I am unable to maintain a fix.


What is your work flow?
Can you post details on your work flow?

1 Like

What I mean
We wanted to setup a base point so

  1. We place the base and log 2 hours of data
  2. We get cors info about 1 before, and an hour after of the collection data. Our nearest cors base is about 13 km form the collection point.
  3. we used the settings posted for ppk processing on the emlid guide
  4. we used cartography to check the location of the base
    Once we get the base coordinates for PPK with a base and a rover
  5. The base was collecting all the time
  6. the rover collected readings for 10 minutes
  7. In RTK Post we set the obs and nav files for the rover, the obs file of the base
  8. In rtk post / options in the position tab we set the base station to lat/long/height. This setting allows to input the coordinates of the base station that we set up in the field with cors data.

You can get a work flow at

he uses multiple points using the survey tooland export them as csv for time stamps for each individual point
we also tried another aproach logging individual ubx files for each point and prostproccesing as described above (not in the video) for each rover ubx.
Hope it helps

Just to a look at the files. You have some reception issues on your rover, as I see it. Lots of cycle slips, which will break your fixes.
Can you post some ground pictures of your environment for this measurement?

Thanks to both of you for taking a look, I appreciate it. Unfortunately I don’t have any images of the environment I can share but it was a petrochemical facility in an open patch of land surrounded by jungle. Let me try to better explain what our workflow was and exactly where I’m having issues.

The goal of this exercise was to use a Base-Rover PPK workflow to geo-locate ground control targets in a remote location. Typically we use just a rover with VRS real-time corrections for this type of work, but this is our first project where real-time corrections were not available. I found an operating CORS network around the area and planned to use that to PPK the base location before processing the rover data, using the nearest reference station that was 50km away. The field work went as such:

1- Set up base receiver on tripod, turn on, and let it log for >1hr
2- Set up rover receiver on bipod, turn on, and let stand still while logging for >10min before starting
3 - Travel to first GCP target and log data on the point for >5min. Use the reachview survey to track the time that the point is occupied
4 - Travel to the additional GCP targets and occupy the point for 1min, again using reachview survey to track the occupied time
5 - After logging at all the points, power down the rover then followed by the receiver

Once the data was logged, the PPK workflow was planned as follows:

1- PPK the Base log against the CORS station to determine an accurate base location (static)
2- PPK the Rover log against the Base log to determine accurate rover pos track
3- Use PPK point extractor to calculate the centroid of the measurements taken while the GCP target was occupied

I’m mainly having issues processing the base log against the reference station to determine it’s position. I’m only getting a 20-30% fix that varies about ±5cm. I am able to successfully process the Rover logs against the Base with ~70% fix, but without an accurate base location I fear that the resulting data will have a global shift equal to the error of the calculated base location.

Hopefully that helps to better explain my issue. I’m guessing the rover cycle slips were caused by the way our field tech carried the equipment from point to point, but I don’t think that should affect much. I was able to successfully calculate the GCP locations using “Average of Single Position” for the base, so if I am able to determine the accurate location I feel confident that our data will yield usable results.

Any ideas on how I can improve the CORS-Base positioning?

Ah, let me take one more look at the base log + cors data then. Thank you for the elaboration.

I am getting the same thing, 35% ish fixes. One thing I noticed was not that many GPS-constellation sats. That coupled with the CORS only providing GPS+GLO, you are not having too many sats for processing a longer(-ish) baseline, even for multifrequency.
I also tried using Fix-and-Hold, but even that does seem to improve the ratio of fixes by a lot. Will try a few more thing, but it’s not looking that good.

Well it’s reassuring that you’re getting the same results. At least with acceptable solutions from the base/rover the local accuracy should remain intact. I will let the client know of the global discrepancy but it shouldn’t be a major issue. Please let me know if you are able to get a coordinate worth using. I will continue to work over my results and just use the best data I can come up with.

Thank you again for the help. This has been my first encounter with PPK and has shown me a lot.

Have you considered submitting the file for PPP with i.e. NRCAN ?

To answer my own question, this is the result of PPP of the Base, using NRCAN Rapid (wait 1 week more to get static, and better precision):

Latitude (+n) Longitude (+e) Ell. Height
ITRF14 (2019.9) 10° 12’ 33.43670" -61° 16’ 26.30939" -3.933 m
Sigmas(95%) 0.009 m 0.023 m 0.042

Attached are the resulting zip with all the files:
full_output (6).zip (3.5 MB)

This topic was automatically closed 100 days after the last reply. New replies are no longer allowed.