Why I preach multiple observations

Matthew D. Sibole

Well-Known Member
5PLS
Anyone who I have sold to or have heard my theory on taking shots in the woods knows that I preach taking multiple observations. DPOS is great but it is not a short cut in "Do I need to take more than one shot on this pin or not?". I will post some screen shots of a job that I am currently working on. The pin in question is under a couple of 18" limbs of a 48" post oak tree. The limbs are about 10' off of the ground.

On Thursday I went out and shot this pin with point number 112 first. Point number 112 did make it all of the way through my strict boundary profile (Confidence Level of 25, Consistency 20, 240 epochs, Validate with 2 minimum engines). The PPK did not agree with the RTK shot on point number 112. It did not take very long to get all of the way through my boundary profile. I was on it for a total of 491 seconds and had 305 seconds between fixes that agreed. My peak to peak horizontal error was 0.144 but mainly just a few outliers causing the such a large spread.

112 screen shot.jpg


I shot the pin again with a point number of 113. This time it did not make it through my boundary settings. I ended up stopping the session at 2080 seconds or 34 minutes. The rtk shots had 117 epochs that agreed over a 136 second time span. You will notice that I put in the point description ?RTK. I do this so I know to review the data when I get back into the office.

113 screen shot.jpg


After I ran things through DPOS that evening point number 113 checked both PPK and RTK within 0.12' in the north and 0.03 in the east and 0.12' in elevation. So for such a bad place I think those two agree pretty well. Especially considering the peak to peak error on the rtk shot was 0.161'. Here is the kicker it does not agree with point number 112 by 0.52' if I hold the PPK data on point number 113.

points list.jpg


113 graph.jpg


I went back out on Friday to verify which shot was good. I got an RTK shot on the pin again. This time it did not meet my boundary profile criteria but it did meet my person criteria to be able to stop it and feel good about the shot (more than 2 epochs that have a time span of at least 240 seconds between fixed epochs). I stopped the shot at 1955 seconds or 32 minutes with only 3 epochs that agreed with each other over a 691 second time span. It agreed with shot 112 from the previous day within 0.083' horizontal and 0.010' vertical.


131 screen shot.jpg




Here is the moral to the story. Do not trust a single shot even if you get a DPOS solution that agrees with it. As surveyors we expect the next surveyor to follow in our footsteps. We need to make sure we can repeat our results so the next guy can repeat them as well!
 

Nate The Surveyor

Well-Known Member
Having played with similar to the above equasion, I will take a stringent rtk shot, (a few epochs, over a long time) ABOVE a ppk shot, over 30 minutes, every time.
2 rtk shots, that agree reasonably well with each other, with an hour between shots, is pretty golden.
Rtk is fundamentally more robust, in challenging places.
 

Shawn Billings

Shawn Billings
5PLS
As I was reading, I was sure that 112 was the good shot compared to 113 until I saw that PPK on 113 agreed with RTK on 113. At that point I was perplexed. I have not observed an RTK observation agreeing with PPK and being wrong, but it's not something I'm as certain of as your RTK settings which passed on 112. This is very good information, Matt. Thank you for sharing. Ultimately the only way to know was that return trip for 131.
 

Matt Johnson

Well-Known Member
5PLS
On Thursday I went out and shot this pin with point number 112 first. Point number 112 did make it all of the way through my strict boundary profile (Confidence Level of 25, Consistency 20, 240 epochs, Validate with 2 minimum engines).

This point does not look like it passed the Validate phase. There should be at 250 epochs if it completed the Validate phase, 240 as required by your settings plus 10 more for the validation.
 

Shawn Billings

Shawn Billings
5PLS
Here is the moral to the story. Do not trust a single shot even if you get a DPOS solution that agrees with it.

This statement is definitely safe, but I don't agree with it completely. Our boundary profile collects the equivalent of many points in one observation compared to other receivers. I would feel very confident in 112 and 131 based on what we know about the Triumph-LS performance under canopy. Point 131 could be considered three different points collected in one observation. Each epoch was based on a new fixed solution and the first and last are separated by 11+ minutes. You don't have a lot of epochs to build the precision, but you do have two different fixes that are separated by 11 minutes which demonstrates good accuracy.

The big news I see is that it is possible to have RTK and PPK agree and for the result to be incorrect. The fact that 113 was wrong is not terribly surprising with only 136 seconds of duration (and probably much less separation between first and last fix), but that the PPK agreed with it and it was wrong is definitely newsworthy.
 

Shawn Billings

Shawn Billings
5PLS
This point does not look like it passed the Validate phase. There should be at 250 epochs if it completed the Validate phase, 240 as required by your settings plus 10 more for the validation.

It's interesting that it shows 241 epochs... That would suggest that the validate engines briefly fixed and were then lost, unless there is some other explanation for the epoch count going beyond 240.
 

Matt Johnson

Well-Known Member
5PLS
It's interesting that it shows 241 epochs... That would suggest that the validate engines briefly fixed and were then lost, unless there is some other explanation for the epoch count going beyond 240.

When 240 epochs were collected a command would have been sent to reset the RTK engines, in this case it appears an extra epoch was recorded before the engines actually reset and the validate phase never fixed.
 

Matthew D. Sibole

Well-Known Member
5PLS
Matt you are correct I did stop the shot after it reached 240 epochs seeing as though I had more than 240 seconds of data. I typically do not care about validate since I repeat the observation anyway.

Shawn, I understand how Jfield handles the collection of data as compared to other receivers. However, 1 stored shot is 1 stored shot regardless of the process needed to store the position. 1 stored position that can not be duplicated is simply not enough regardless of the processes used.

With all of that said it is just my opinion. That is what is so great about the LS and Jfield. It allows users to make those choices for themselves.
 

Matt Johnson

Well-Known Member
5PLS
However, 1 stored shot is 1 stored shot regardless of the process needed to store the position. 1 stored position that can not be duplicated is simply not enough regardless of the processes used.

If the validate phase would have been allowed to complete it would be very unlikely that this group would have passed and and the point would not be accepted and stored with the those coordinates.
 

Matthew D. Sibole

Well-Known Member
5PLS
If the validate phase would have been allowed to complete it would be very unlikely that this group would have passed and and the point would not be accepted and stored with the those coordinates.
I know that point 113 was very skeptical as noted in my ?rtk description posted. I had little doubt that 112 was bad due to the amount of time between fixed epochs that agree which is what you are eluding to in your suggestion. Waiting the addition time on 112 I think would have been wasted time. I would rather accept the shot and start another one.
 

Nate The Surveyor

Well-Known Member
I tend to agree with Matt's statement above.
It's a system designed to give true answers in the field.
Use validate for field confidence.
N
 

Duane Frymire

Active Member
I know that point 113 was very skeptical as noted in my ?rtk description posted. I had little doubt that 112 was bad due to the amount of time between fixed epochs that agree which is what you are eluding to in your suggestion. Waiting the addition time on 112 I think would have been wasted time. I would rather accept the shot and start another one.
What was the difference between rtk and ppk on point 112?
I too have seen rtk and ppk agree in situation similar to point 113. But I don't always take a second shot I've become so confident in the last validate stage. It can take awhile but I've never had a bad position if it does.
 

Nate The Surveyor

Well-Known Member
I could talk for a long time about the things I learned with L1 only static receivers.
This is for woods work. And, heavy canopy.
But, to make it short, rtk, with resets, over long time, is going to be stronger, (as in no big busts, or errors) than static. Unless, someone comes up with some great new changes in post process gps. Rtk can do verify, and all, 2x, and process up 2 shots that are 0.20' apart.
So, I argue, use rtk to get the right initial location. Then, use multiple observations to refine, and look at shot skatter. Looking at the screen and watch out for long "string shots". These ones are usually walking and generate piles of coords that have some sort of elipse 1/2" wide x 4-5" long,
If I were doing the job, and saw results like @Matt Sibole was seeing, I'd assume in the field that one, complete, with verify, was good, and that 113 was bad. Planning a 3rd observation would be started, as soon as I saw the 0.52' discrepancy. His hope that ppk from 113 would agree with rtk 112, is not very strong.
Point is, 2 rtk shots, with Matt's settings, and verify, that agree well, is what we should be looking for...
 

Matt Johnson

Well-Known Member
5PLS
But, to make it short, rtk, with resets, over long time, is going to be stronger, (as in no big busts, or errors) than static.

This should not be true. Post processed static solutions should be better provide the observation is long enough as it has the ability to look at all data at the same time and process the data forwards and reverse. At some point in the near future I plan to do a test to compare RTK solutions vs Justin post processed solutions.
 

Nate The Surveyor

Well-Known Member
Well, my statement is backed up with many hours of testing, and expiramenting. It's not theory.
Although what Matt says SHOULD be true, it is NOT true, by my experience, in the woods, in the dirt.
Post process gps was built for good answers, in the field, not woods. Pushing it into the canopy may require a re-write.
It's not there yet. (No offense to you Matt, and you are a good man, but I had to test and solve until I knew.) It's what surveyors do, who are born this way!


Nate
 

Sdrake14

Active Member
I could talk for a long time about the things I learned with L1 only static receivers.
This is for woods work. And, heavy canopy.
But, to make it short, rtk, with resets, over long time, is going to be stronger, (as in no big busts, or errors) than static. Unless, someone comes up with some great new changes in post process gps. Rtk can do verify, and all, 2x, and process up 2 shots that are 0.20' apart.
So, I argue, use rtk to get the right initial location. Then, use multiple observations to refine, and look at shot skatter. Looking at the screen and watch out for long "string shots". These ones are usually walking and generate piles of coords that have some sort of elipse 1/2" wide x 4-5" long,
If I were doing the job, and saw results like @Matt Sibole was seeing, I'd assume in the field that one, complete, with verify, was good, and that 113 was bad. Planning a 3rd observation would be started, as soon as I saw the 0.52' discrepancy. His hope that ppk from 113 would agree with rtk 112, is not very strong.
Point is, 2 rtk shots, with Matt's settings, and verify, that agree well, is what we should be looking for...

Right on Nate! This is good ole horse sense....
 

Sdrake14

Active Member
Well, my statement is backed up with many hours of testing, and expiramenting. It's not theory.
Although what Matt says SHOULD be true, it is NOT true, by my experience, in the woods, in the dirt.
Post process gps was built for good answers, in the field, not woods. Pushing it into the canopy may require a re-write.
It's not there yet. (No offense to you Matt, and you are a good man, but I had to test and solve until I knew.) It's what surveyors do, who are born this way!


Nate

I agree is probably one of the smartest surveyors I ever talked to. But I have to admit the time in the field trying/testing your gear and proving it is where I come from too. Trust but verify. I used to think (and based on GPS early statis quo) that static post processing was superior but what I saw in practice proved otherwise, and particularly since I used Javad's technology. I might think different with other vendor's junk Nate since most, in my practice, do not offer you the real time measurement quality feedback you have with Jfield.

I spent nine months using the Triumph 2 to slope stake 3.5 miles of mountain pass road in northern California old growth woods and by allowing the tools in Jfield to do it's work I can honestly say that I got a low percentage of bad shots and when I did I knew it it 99% of the time when I walked away from that shot in the field. I was able to then use PPK, post processing (beast mode) to further analyse suspect shots. I am empowered to know when I need to return or choose a different method to produce this measurement. In some cases returning was a painful or prohibitive proposition and having the feedback on the spot allowed me also to decide to stay and take a bunch of shots to have a statistical dice roll on a good one, and in Matt's defense I will say on some of these at the end of the day the only good (or acceptable) shot I had was a PPK one.

Ther are no rules except repeatablity.
 
Top