Fast average of BASE position in RTK

While in the field, we were not able to get at reliable connection to a CORS network. We were forced to average our Base Station in SINGLE mode for about 15 minutes. This wasn’t a big deal because we really only needed relative accruacy between points taken on the rover. However, I am concerned that even relative accuracy will be effected if the base position is off significantly.

My question is; if I put in Base coordinates in WGS84 but if they are wrong by over a meter or two will this introduce an error in my relative measurements? I understand that this is normal workflow, but I would like to have an idea if there is some inherent error that is being introduced into the relative accuracy, even if only a little. Also, I am wondering how much error can a base setup tolerate before it starts effecting FIX times etc. (I know that if the base coords are off by many Km then you will not get fix and the corrections are not useful).

1 Like

They surveyors will know the best, and should chime in.

The difference between NAD83 and WGS84 in my area is a couple of meters, I can set the coordinates in as either without issue to use either system. If you long average your base you will still most likely be a meter out. So I think you will be precise to your own measurements.

The challenge is coming back to do more relative measurements, you will need the exact base coordinates you used and the exact mount point again.


“My question is; if I put in Base coordinates in WGS84 but if they are wrong by over a meter or two will this introduce an error in my relative measurements?”

Well yes, but not at a level that matters for almost any application. The error will be well below a millimeter for typical baseline lengths. This is done all the time even for “high end” surveying applications where, say NAD83 coordinates are entered for the base and though they are different than WGS84 coordinates by a meter or two - and the software may expect WGS84 coordinates - the result is still a high accuracy NAD83 coordinate.

You can work though the math and the extreme distances to the sats (~20,000 km) versus the offset distances you are talking about (1-2 meters) should give you an idea that the 1-2 meter offset won’t matter.

You can also try a few experiments. Process a baseline and change the base coordinates by 1 meter, then 10 meters, then 100 meters and look at the relative baseline length change. It’s not going to be “significant”.

Now depending on the software you might not be able to get the base coordinates really wrong as sometimes there is a “sanity” check where the autonomous solution is used and if the error is too big it will complain. Some software might flag things if you are say 50 meters out. Last time I tried this with an earlier version of RTKLIB I think it let me enter base station coordinates a kilometer out.


No, this is common practice even among surveyors. You are still within the RTK specs in relative in this situation.


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