Measuring twice the same points give me different positions

Hello,

Sorry for the beginner question, but this one is puzzling me.

We are using Reach 2 Base + Rover. I measure twice the same set of six points, moving the base (and restarting) between the two acquisitions.
The points are recorded in FIX mode.

My problem is that there is quite a difference between the two measures (Easting is pretty ok, but Northing shows a delta of almost one meter). RMS and everything looks just fine

I frankly do not understand what I’m doing wrong.

Here is a plot for the error between the two measure of the same point, for each six points
And below are the export .CSV from ReachView App

Thanks a lot for your help,

Alex

set1.csv

Name,Easting,Northing,Elevation,Description,Longitude,Latitude,Ellipsoidal height,Easting RMS,Northing RMS,Elevation RMS,Lateral RMS,Antenna height,Antenna height units,Solution status,Averaging start,Averaging end,Samples,PDOP,Base easting,Base northing,Base elevation,Base longitude,Base latitude,Base ellipsoidal height,Baseline,CS name
1,696528.648,4138403.025,56.167,,23.21948061,37.37155886,56.167,0.011,0.010,0.010,0.015,0.534,m,FIX,2022-08-14 21:22:14.0 UTC+03:00,2022-08-14 21:22:19.0 UTC+03:00,26,1.6,696530.395,4138393.699,57.482,23.21949785,37.37147449,57.482,9.515,WGS 84 / UTM zone 34N
2,696529.971,4138403.037,56.172,,23.21949554,37.37155869,56.172,0.012,0.011,0.010,0.016,0.534,m,FIX,2022-08-14 21:22:27.2 UTC+03:00,2022-08-14 21:22:32.2 UTC+03:00,26,1.4,696530.395,4138393.699,57.482,23.21949785,37.37147449,57.482,9.378,WGS 84 / UTM zone 34N
3,696534.137,4138403.118,56.172,,23.21954257,37.37155853,56.172,0.012,0.010,0.012,0.015,0.534,m,FIX,2022-08-14 21:22:42.4 UTC+03:00,2022-08-14 21:22:47.4 UTC+03:00,26,1.2,696530.395,4138393.699,57.482,23.21949785,37.37147449,57.482,10.164,WGS 84 / UTM zone 34N
4,696538.248,4138403.174,56.174,,23.21958898,37.37155817,56.174,0.010,0.010,0.011,0.015,0.534,m,FIX,2022-08-14 21:23:00.6 UTC+03:00,2022-08-14 21:23:05.6 UTC+03:00,26,1.3,696530.395,4138393.699,57.482,23.21949785,37.37147449,57.482,12.330,WGS 84 / UTM zone 34N
5,696534.199,4138397.597,56.160,,23.21954181,37.37150879,56.160,0.010,0.010,0.012,0.015,0.534,m,FIX,2022-08-14 21:23:18.4 UTC+03:00,2022-08-14 21:23:23.4 UTC+03:00,26,1.2,696530.395,4138393.699,57.482,23.21949785,37.37147449,57.482,5.505,WGS 84 / UTM zone 34N
6,696530.062,4138397.537,56.147,,23.21949511,37.37150912,56.147,0.011,0.010,0.011,0.015,0.534,m,FIX,2022-08-14 21:23:34.6 UTC+03:00,2022-08-14 21:23:39.6 UTC+03:00,26,1.4,696530.395,4138393.699,57.482,23.21949785,37.37147449,57.482,3.932,WGS 84 / UTM zone 34N

set2.csv

Name,Easting,Northing,Elevation,Description,Longitude,Latitude,Ellipsoidal height,Easting RMS,Northing RMS,Elevation RMS,Lateral RMS,Antenna height,Antenna height units,Solution status,Averaging start,Averaging end,Samples,PDOP,Base easting,Base northing,Base elevation,Base longitude,Base latitude,Base ellipsoidal height,Baseline,CS name
1,696528.635,4138403.984,51.349,,23.21948071,37.37156750,51.349,0.012,0.012,0.011,0.017,0.534,m,FIX,2022-08-14 22:53:59.0 UTC+03:00,2022-08-14 22:54:04.0 UTC+03:00,26,1.3,696537.259,4138407.122,52.714,23.21957887,37.37159394,52.714,9.220,WGS 84 / UTM zone 34N
2,696529.950,4138403.996,51.355,,23.21949556,37.37156732,51.355,0.013,0.011,0.012,0.017,0.534,m,FIX,2022-08-14 22:54:16.2 UTC+03:00,2022-08-14 22:54:21.2 UTC+03:00,26,1.2,696537.259,4138407.122,52.714,23.21957887,37.37159394,52.714,8.003,WGS 84 / UTM zone 34N
3,696534.128,4138404.081,51.370,,23.21954273,37.37156720,51.370,0.010,0.011,0.011,0.015,0.534,m,FIX,2022-08-14 22:54:40.6 UTC+03:00,2022-08-14 22:54:45.6 UTC+03:00,26,1.1,696537.259,4138407.122,52.714,23.21957887,37.37159394,52.714,4.437,WGS 84 / UTM zone 34N
4,696538.250,4138404.111,51.359,,23.21958926,37.37156660,51.359,0.010,0.010,0.012,0.015,0.534,m,FIX,2022-08-14 22:55:02.4 UTC+03:00,2022-08-14 22:55:07.4 UTC+03:00,26,1.1,696537.259,4138407.122,52.714,23.21957887,37.37159394,52.714,3.276,WGS 84 / UTM zone 34N
5,696534.166,4138398.578,51.335,,23.21954170,37.37151764,51.335,0.011,0.011,0.012,0.015,0.534,m,FIX,2022-08-14 22:55:24.8 UTC+03:00,2022-08-14 22:55:29.8 UTC+03:00,26,1.1,696537.259,4138407.122,52.714,23.21957887,37.37159394,52.714,9.129,WGS 84 / UTM zone 34N
6,696530.043,4138398.509,51.329,,23.21949516,37.37151789,51.329,0.010,0.010,0.013,0.014,0.534,m,FIX,2022-08-14 22:55:45.0 UTC+03:00,2022-08-14 22:55:50.0 UTC+03:00,26,1.2,696537.259,4138407.122,52.714,23.21957887,37.37159394,52.714,11.270,WGS 84 / UTM zone 34N

when you move the base, that base has to be in the same coordinate system. Your query is not very clear. You could use shift to base to take a known point and calculate the displacement and everything would stay in the same system

How much did you move the base? What solution did you use to collect the base coordinates?

I see that you only had 26 samples? What was your duration? Rovers normally run at 5Hz which is only 5 seconds.

1 Like

My main concern is how you achieved the broadcast base position. Did you accumulate and average a SINGLE solution for the base in both sessions?

3 Likes

Hi,

Thanks for all your answers. And sorry, Agrimgalina, for not being clear enough.

Your advice led to a new set of measures.
Same protocol, but more repetitive:

  • Turn on base and wait for 20 minutes.
  • Turn on rover and wait to be in FIX mode.
  • From ReachView App, start a new project (UTM 32N EPSG 32634, averaging over 1 second). We average over 1 second, because our “real” field is in water, and it’s a real mess to keep the pod vertical for a few seconds while swimming in 2.5m. However, a few centimeters’ precision is perfect for our needs.
  • Record the same 6 points (on a tennis court)

The three data sets for the surveyed points still show good precision, but poor accuracy (.7m easting, 1.6m northing).

The base station was moved three times on a tennis court (5-10 meters apart).
My first error was to think that the base station was able to average an accurate position.

So, I tried averaging the base station position (over the 20 minutes of the experiment), taking the points where the sde and sdn are <=12cm

And there, we see that there is some coherence in the shifted positions.

But the shift between the base and the survey does not align…

My fear is that we messed up our multiple-day survey, by recording points every day with the base station not exactly at the same position and not taking reference points.

What is the way to align to a better accuracy ?
Is our solution to post-process the base RINEX with an NTRIP station (we are in Greece)?

Thanks again for your help,
Alex

And just to be sure I did not messed up (also) the survey and .llh files, we QC the time intervals

Yes, it does average single in two bases. The measurements will never be the same because they are not in the same system.

1 Like

Averaging the base position and comparing was dull… because the base position has actually changed…

Nonetheless, I do not actually understand if I have a solution to my accuracy problem.
Alex

If you cannot return to the site and survey your base points again, not really.

Normally, the simplest way would be to put a semi-permanent marker in place, average a position for that point on the first day, note the coordinates down and for each subsequent session, place your base over that point and enter the recorded coordinates manually (or use the new feature where you can use a surveyed point in your project as the base position). Your absolute accuracy will not be the best, but at least it will be consistent between each session.

3 Likes

Hi Alexandre,

In ReachView 3, there are several options to set up the base position. The default one is Average Single mode.

Single solution is a few-meters accurate. So, your base position is shifted from its true position on the Earth. Points you take with the rover are still centimeter-level accurate, but in relation to the base.

When you reboot the base, it re-averages its position in Single again. And as accuracy of Single solution is low, it won’t be the same coordinates even if you placed the base exactly over the same point on the ground.

If you want to match your measurements from different days, you need to write down the averaged coordinates and use them for all future surveys. But note that all your coordinates will have relative accuracy: they will be shifted from their real position on the Earth.

I’d suggest you check this article about relative and absolute positioning. Probably, you will decide that you need an absolute positioning.

Did you record RINEX during your surveys?

4 Likes

Thank you Svetlana for your clear answer.

And yes, we do have RINEX

Alex

Averaging for 1 second? Your points report shows 26 samples. That either a 26 second at 1Hz or 5.something seconds at 5Hz.

Sounds like you might benefit from a bipod attachment for your rover.

1 Like

Thanks for the bipod idea. I don’t know how we could handle that in the sea, but that is worth some thought.
And the 26 measures were coming from the first set. On the three last one, we were at 6 measures (1s at 5Hz)

Alex

using a bipod is not the solution, there will never be a difference of one meter. You have to study a bit of reference systems. I have been using RS2 for years and the differences depending on the method is not more than 1 cm.

1 Like

Great!

Then, as you said, you can request logs from the nearest NTRIP reference station and post-process them with the base RINEX.

In the CSV, you have base coordinates that were averaged. You will need to calculate the shift between averaged and post-processed base position. Then, just apply this shift to surveyed points.

2 Likes

I recommended a bipod or some type of support because he said that it was hard to keep the rod level which is why they were only shooting for one second. I didn’t say it was the solution as it is obviously everything else that went on. There’s no way you’re going to get repeatable accuracy with that duration.

Thanks for all your ideas and comment.

The bipod and the post referencing are two different dimensions.
And we’ll try them all.

We are on a huge survey/photogrammetry challenge in shallow water in Greece and discovering new dimensions of tackling problems.

Alex

1 Like

Thank you Svetlana, you nailed it :wink:

And we’ll be more cautious with the base in the future.

Alex

1 Like

You’re welcome! It was really good decision to record RINEX in the field. Now you don’t need to redo the job :slightly_smiling_face:

1 Like

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