GPS Processing Quesion

Typically more redundancy is better. If one of your control points you occupied with a base is of lesser quality (ie obstructed sky or lower order of accuracy) it might make sense to omit one. Otherwise leave both in. The least squares adjustment will take into account the estimated quality of each vector and weight them for the best position of your unknown point. This assumes both vectors are fixed with no appearance of vector quality issues.

If I understand you, you have two vectors from a known point to an unknown point you want to determine. In that case I will usually use both vectors and adjust. Understand that the two vectors are never of identical quality. Adjustment software, such as contained in TGO, will weight the two vectors differently according to the vector quality information (ie/ the covariance matrix) and adjust appropriately.

If you mean that you have vectors from two known points to an unknown the same applies, except you have the added twist that your two known points have standard errors, which also play into an adjustment. So take advantage of the software's capabilities.

That said, checking the result of that adjustment against the results of an unadjusted measurement is always prudent.

In brief, I like to incorporate all the data I collect unless I can show that there is a blunder.

Yes I have 2 different bases & 1 unknown point for observation. I do not understand all of those dx,dy,dz errors etc.

So if I understand correctly, I just use all observe data & resulting point coordinates is the Best solution available?
Thank you everyone.

> So if I understand correctly, I just use all observe data & resulting point coordinates is the Best solution available?
> Thank you everyone.

Probably. Assuming that the two possible unadjusted results are not dramatically different - which would suggest a blunder somewhere- the results of all the data combined would likely be best.

Double redundancy is a plus when rover points are close together.

I do most of my static using two base stations and process with GNSS Solutions.

I get a fixed solution on one from CORS or TxDot station/s.

Process a solution for base number two.

Change both base stations to stationary control points.

Process all the rover setups based upon the location of the two bases.

When the results are within tolerances I am ready to download and COGO.

When the rover points are several miles apart, I leapfrog units around in pattern that resembles a triangular pinwheel mesh. Download everything together and process everything together.

0.02

Different softwares different results?

Just downloaded GNSS trial version based on previous post recommendation.
Maybe I am not yet familiar with the settings but is it common to have different resulting coordinates from different post processing software? I used default settings for accuracy filter - 0.01 for horizontal & 0.02 for vertical limits.

I think difference is less than 2-3 cm which is small for a vector of 8 km.

But shouldn't same GPS data yield same coordinates?

If not, what is the defacto post processing software that you would recommend?
Or should I just stick to the software of my GPS brand?

Adjust

Once you have your adjusted point, you hold it fixed in your software and release constraints on your control points. Now adjust again and compare your solution control coordinates with the published coordinates. That tells you in a nutshell what your software did.

In essence that is what RSGPS the software within our OPUS-RS does using up to nine control points.

Paul in PA

Different softwares different results?

Howdy,

For what it's worth, the fact that different software packages yield different (though small magnitude) results is unsurprising. Processing GPS vectors involved the choices in the models to apply and how they are applied. Some tools are optimized for ease of use and the expectation that work will be done over short baselines making some corrections unnecessary.

Adjustment software should yield close agreement if the tool uses all available data. Some packages, for example, do not use all elements in a variance-covariance matrix instead using only the standard deviations of the baseline elements. In other words, they use the standard deviations of the DX, DY and DZ but not the correlations of the DXDY, DXDZ and DYDZ. Hope that makes sense.

On a related note, I recently submitted a 13-hour GPS file to all the automated free GPS processing sites. The link below is to a table of the results and some basic comparisons. Interestingly some tools use differencing and others point positioning. As can be see in the comparisons, good agreement was obtained. What is the correct coordinate for the "new" site?

http://geodesyattamucc.pbworks.com/w/file/fetch/64369423/GPSsolutionsCompared_2013.pdf

Good luck,

DMM

Adjust

:good:

Different softwares different results?

I agree with Mike, with one added caveat. You will see differences because different processors use different techniques and estimates to arrive at a result, however there are also user assigned differences that can make a difference, particularly the mask angle and perhaps less so, the reference satellite, which the software will generally select the highest satellite in the sky.

Atmospheric delays aren't exact, they're estimated. Very closely estimated, but still estimated.

Different softwares different results?

As Mr. Billings notes, users are frequently the source of bad solutions. Just because you can change a default does not mean you should.

As I understand it, the development of OPUS was in large part an attempt to make solutions consistent and rigorous and repeatable. It achieved this by applying a uniform approach to processing.

BTW, the web site in my earlier message includes copies of the solutions from each of the tools.

Hope you're not using TGO, the processor timed out some year and a half ago at 999 999 999 seconds. Much less processing with TGO since then.

Same software different results?

Some processors can't make up their own minds. I can take a bunch of static sideshots, fix the base coords, and push the process button over and over and watch the coords of the sideshots change every time. (Topcon Tools) Trimble didn't do that.