RTK vs PPK

[Joe Paulin]

Question for the experts: Using base/rover rtk in challenging canopy conditions, and collecting PPK data as well, which is more reliable/accurate assuming the PPK session is over 10 minutes?

Looking at the point data comparison in the LS, the HRMS and VRMS is almost always lower for the PPK solution, making it seem like the better choice to select vs the RTK solution. I am talking about where the solution differences are small, under a tenth. Just looking to export the BEST coordinate. When I choose auto-select in the LS, it usually chooses the RTK solution.

Thanks!

 

[Nate The Surveyor]

Well, Joe, I personally have wrangled with such questions in life. It's nearly a philosophical question. I used to ask my dad, "If you push the gas all the way, and the brakes all the way, which one will win?". To which, he used to laugh, and say, well, it all depends on the condition of the repair, of the engine, or brakes at any given time! It was an old '57 Ford Panel truck....

So, I have played with the above equation, and have developed some ideas.... nothing for sure, but some ideas.
In super challenging places, generally, I think RTK is MORE reliable. But, there have to be exceptions.

So, I have thought about it. Why not use a total station, (Based on long PPK solutions) to get the "real coordinate", and then sit on it all day, and then ppk it, and then look at all the data.....
This is what I sort of think.....
USE the rtk shots alone, and AVG them. And, it will give you one position.
USE PPK alone, and AVG them, and it will give you another.
I think the 2 above observations, (avg derived from RTK, and AVG derived from PPK) that there will be LARGER coordinate piles from PPK, than from RTK.
Meaning this. LARGER scatter, from PPK, than from RTK.
For instance, I did 3 PPK shots some months ago. 2 agreed real well, and another was some 0.2' away. RTK was NOT available there.

So, here is my condensed thoughts.

PPK requires approximately 1.3 to 1.5x time (Time = data) for the same quality of solution, as RTK.

So, IF you allow extra time, average BOTH solutions (RTK, and PPK) into their own piles, they will eventually become the same coord, position, (+ - 0.015')

Averaging RTK solutions, will give you screen shots, (how big the pile is), but, PPK does not give this...


So, IF I had 5 RTK shots, that took 4 hrs, and piggy backed onto this, 5 ppk shots, over the same time period, I'd feel that the PPK shots were weaker, by some magnitude of 1.33 for a multiplier.

OR you could say it this way. IF a RTK shot takes 15 minutes, to get a shot, that looks pretty good, then it would take a 20 minute PPK shot, to get the same quality.

This is based upon my thoroughly non scientific, experience.

Having said that... I have wished for a mechanism to slightly weight the RTK shots, with a PPK leaning.... But, I am not sure how this would work out.

Thanks for bringing this discussion up. It would be fun to have MORE data to analyze this question. But, it is on my mind....

Hey, ain't it great to be fussing about a few hunnerts? This whole conversation could not occur, with any other species of GPS....

N

PS, I'm NOT an expert... I am a man who suffers from curiosity...

 

[Matt Johnson]

All the research tools needed to answer these type of questions are available in JField. To compare the quality of the RTK solutions to those from DPOS I would suggest the following procedure:

1. Set the base up in an open area with it aligned to north (DPOS processes with azimuth dependent phase center variations).
2. Set the LS up in test location with it aligned to north. Adjust the height of the LS to be the same as the base or measure the height difference the two receivers with an optical level so that position accuracies, rather than just precisions can be computed later.
3. Set the Minimum Duration to some time that you wish to test that is sufficient for DPOS base-rover processing.
4. Set Auto Accept to Any.
5. Set Auto Re-Start to Always
6. Let the LS collect many observations of the same point
7. Compare the RTK results to the PPK solutions by selecting all RTK solutions and then all PPK solutions. We have "Survey Statistics" for this. Average the results of the points and select it as a reference point and then edit the height of this point so that it corresponds what was measured with your optical level. Next select that point as a reference point for the Distance Filter:
   
   
   
   
   
   Then select a point in the points list and choose "Statistics" from the context menu. You will be presented with a screen like this: