Beast v. DPOS, and the Greatest is...

D

Duane Frymire

Active Member
Both, working together.

I have what appears to be a good RTK solution, but the PPK was selected by the new software version. I talked with John Evers to get the profile set as Shawn mentions below so that it always collects raw data for PPK. There was a couple of feet difference in the solutions, so one is just bad. I checked with the total station and it turns out to be a bad RTK shot, and a good PPK. The PPK is base corrected. I am not collecting CORS corrected because stations are 25 miles away and base corrected would presumably always be better.

Anyway, John asked me to post the files here for the development team to look at. Point #21 is the point in question.

File is too large, so sending to Vlad via email.

Duane
 
D

Duane Frymire

Active Member
Matt Johnson said:
What collection settings did you use for the RTK point? Did you use Verify with Resets and Validate?
Click to expand...
Yes, Verify with resets minimum 2 engines, Validate 1, had 130 epochs, showed very good rms as well. Tried to post a screen shot but didn't work. Alexy verified the PPK solution is the correct one. Not sure what went wrong with the RTK one, maybe they can figure it out.
 
Darren Clemons

Darren Clemons

Well-Known Member
Duane Frymire said:
Yes, Verify with resets minimum 2 engines, Validate 1, had 130 epochs, showed very good rms as well. Tried to post a screen shot but didn't work. Alexy verified the PPK solution is the correct one. Not sure what went wrong with the RTK one, maybe they can figure it out.
Click to expand...
I also have validate with a minumum of 2 engines. I won't let it do anything to do with an RTK shot set to only one engine. The "one" engine fix can and usually is engine one in coverage. I have seen a slight few bad RTK fixes (maybe 4 in 6 months) with 120 epochs and 10 more with validate w/2 enigines but like Matt says it can do that in 30 seconds. If you set minimum epochs to at least 600 (5Hz rate for 120 seconds) and validate with min 2 engines the chances of that storing a "bad" fix becomes extremely unlikely. Kinda like Bigfoot - not sure anyone's actually seen one yet!
 
Nate The Surveyor

Nate The Surveyor

Well-Known Member
Having been surveying for a long time, my "base line score" that I compare against is how long it would take with a 2 man crew, transit, and tape.
Getting in a hurry, will allow a lie the sneak through.

So, here is the NATE Safe setting:
Start with Start button.
Stop after 100 epochs
Min Duration is 200 This is critical. It this lets the sats arc across the sky, for an adequate arc to catch the lies. I believe others use 120 or 130 for this, but, I am at the EDGE of the arena, and so I want 100% confidence.
Verify 2
V6 reset YES
Confidence Level 30
Consistency 30
Flashlight Blink YES
Max Groups 10
Validate with at least 2
 
Nate The Surveyor

Nate The Surveyor

Well-Known Member
Yes, David, in the bad spots. I have had mine kick up to 20 on consistency, and then throw it all out, due to 10 groups not being enough!
I think something is in the wings on this, to allow it to delete the worst groups, and keep going.
N
 
D

Duane Frymire

Active Member
Matt Johnson said:
What was your correction rate? You need to be collecting points with a duration of at least 2 minutes in bad areas, 3 minutes is preferred. At 5 Hz 130 epochs could be collected in less than 30 seconds.
Click to expand...
I had 103 seconds, 9.5 confidence, 11 consist. The PPK was 6 min. 10 sec. In other words, I knew there might be a problem with the RTK because only two engines fixed during the process and I only had validate set to one. So went a bit longer for PPK to see how it would compare. I suppose the truer comparison would be the same time for each, but that would discount any speed gains given by 5hz.

I had been using 5 groups but for this exercise went to 3 groups, theory being it will reset tracking sooner and not build time on a bad signal and then reset, as mentioned by Nate.

This was a job I normally would not bother with GPS because time on point would make it take too long. But wanted to test RTK v PPK to see what it would catch, and what would process at reasonable occupation times for the type of project. Generally speaking, the PPK at 5-6 minutes was not adequate either in bad spots where RTK wouldn't fix 2 engines. Errors of 0.3' to 3'.

I'm going to reoccupy some of these with longer times/repeats and see if/when a good solution turns up.

But I'm thrilled to have this comparison of PPK to RTK. So far I haven't found a bad position when the two are in substantial agreement. Really an exceptional tool for analysis of the quality of the data.
 
Adam

Adam

Well-Known Member
5PLS
I have found magical things happen at 10 minutes plus with PPK. I was localizing to a survey with random coordinates, Monday. It was really heavy canopy in the mountains and I wasnt able to get but a few epochs RTK on the points. After processing static rover files the largest descrepancy between desing points and survey was .07' except for one that had no sky view at all. I was seeing D-SNR values of 7-even 10 on the one with no sky view at all.
 
D

Duane Frymire

Active Member
Adam said:
I have found magical things happen at 10 minutes plus with PPK. I was localizing to a survey with random coordinates, Monday. It was really heavy canopy in the mountains and I wasnt able to get but a few epochs RTK on the points. After processing static rover files the largest descrepancy between desing points and survey was .07' except for one that had no sky view at all. I was seeing D-SNR values of 7-even 10 on the one with no sky view at all.
Click to expand...
Thanks, I'll try that. I was doing 20min to half hour on those types with opus rapid static. Seems reasonable that base corrected PPK could produce similar results in less time.
 
Adam

Adam

Well-Known Member
5PLS
Really, 20 to 30 minutes is not bad considering the time it would take to traverse to some of these bad areas.
 
D

Duane Frymire

Active Member
Adam said:
Really, 20 to 30 minutes is not bad considering the time it would take to traverse to some of these bad areas.
Click to expand...
Yes, almost magical really. I'm playing with it on a small lake front lot survey where all can be done with a total station very quickly. Just trying to see how I might use GPS in the most productive way on a job like this.
 
You must log in or register to reply here.
Top