New Guy dumb question #1
- Thread starter jahalnon
- Start date
Been experimenting a lot more with this and I seem to be seeing the following:
Thanks,
John
- with decent sky and minimum multipath, DPOS solutions are consistent with RTK solutions (I was expecting this)
- with moderate canopy and multipath, DPOS solutions were scattered in an unusable cluster. But multiple RTK solutions worked well and were usable.
- with typical winter middle of oak forest canopy and multipath, DPOS solutions were unusable and multiple RTK solutions had mixed results; sort of the usual GPS and RTK caveats come into play and I have to be careful of the results. A bit of a damper on my initial enthusiasm of the LS since it appeared to be giving me good solutions.
Thanks,
John
Adam is right. There are many times I find DPOS works where RTK cannot in really thick canopy, but you have to have enough time in it. Generally, I get good DPOS results in light to moderate canopy with 3 minutes of data. If I'm under canopy where RTK struggles to get a solution, I may need much more time 10-30 minutes. If I am unable to get an RTK solution and I'm relying on DPOS for the position, I will repeat the observation so that I can verify my DPOS solution with a second DPOS solution as it is possible to get a bad fix from DPOS.
Nate The Surveyor
Well-Known Member
Shawn is 100% correct above. The big magic of Javad HAS been a highly sophisticated and refined system, to weed out the bad, while in a multipath environment, with RTK.
So, when you use PPK (That's Post Processed), the game changes. You should use 2 or more observations, to assist in weeding out the bad data. Often, if you have 5-10 minutes worth of data, AND a good RTK shot, with it, the PPK shot also has enough data.
Invest the time, and energy in comprehending what is happening, in this area. Not just theory, but practice.
I came into this, with:
Single channel, post processed GPS.
Then, dual, rtk, and some post processed.
Now, it's all in one. I like it, yes I do. And, Javad keeps refining it. I'll admit, I have been pushing the Javad team some. But, this is the FINEST GPS in the history of the planet. It is the funnest, to watch it flush out a shot, where NOTHING else can go.
Makes my hair stand up on the back of my neck!
N
So, when you use PPK (That's Post Processed), the game changes. You should use 2 or more observations, to assist in weeding out the bad data. Often, if you have 5-10 minutes worth of data, AND a good RTK shot, with it, the PPK shot also has enough data.
Invest the time, and energy in comprehending what is happening, in this area. Not just theory, but practice.
I came into this, with:
Single channel, post processed GPS.
Then, dual, rtk, and some post processed.
Now, it's all in one. I like it, yes I do. And, Javad keeps refining it. I'll admit, I have been pushing the Javad team some. But, this is the FINEST GPS in the history of the planet. It is the funnest, to watch it flush out a shot, where NOTHING else can go.
Makes my hair stand up on the back of my neck!
N
Darren Clemons
Well-Known Member
Everyone has given you great advice above John. It's all about spending enough time in those "tough" areas. If you do the LS WILL, repeat WILL almost certainly get you a shot you can trust. We had to experiment with PPK quite a bit too when we first started using it but with others advice found the 15-20 minute mark in heavy canopy seems to be the magic number. And as Shawn says, sometimes even a bit more.
What my standard has become as far as taking a second 15 minute PPK depends on what RTK I stored when the first 15 minute session completed. I like to get a full validated verified three phase shot but that's not always the case. With the awesome tweaks in the parameters of the groups nothing ever gets reset now so I could (and often do) have a "lead" group of 3,275,2 held as the RTK when my session is complete. If I see that, I'm 99.9% sure that is a good RTK shot. In fact, any time separation over 150 seconds is usually good to go. Also, another trick is if you happened to have stored any points close to where you're struggling to get a shot, always watch the "distance to last" white box. It may give you a very good idea of both distance, but more noticeable, the elevation from the last point stored. If I shot a point 150' feet back and I am on relatively level terrain and my "lead" group is showing me a -15' of vertical I know it's wrong. So many ways to check and keep up with your data on the LS.
Now, what I'll typically see when I process the PPK against this minimal amount of RTK data is a bit more variance than a full three phase validated RTK. Say, something like 0.1'h and 0.15'v, maybe even a bit more, whereas a fully validated RTK against a 15 minute PPK rarely shows more than 0.04' in either h or v. That being said, if I'm in a spot where the Javad can't get full RTK in 15 minutes, 0.1' is as good as gold! It just depends on where you are and what kind of precision you want.
Nate The Surveyor
Well-Known Member
Btw, rtk gps, is a form of distance/distance intersection, from satellites. (you already knew that). So, when a point you wish to observe, falls under extreme canopy, 2 events occur.
1.) Accuracy is degraded, by a few hundredths. And, sometimes a tenth, and even up to 0.16' The worst I've experienced was 0.14'.
2.) There are only certian times of the day, when real bad places are observable.
So... Certian places may not be suitable for gps. They can yield large error ellipses, and take excessive amounts of time.
Places that come to mind, where this might be true, is between tall buildings, up against large structures, inside buildings. Under external stairs.
There simply is not enough data getting through, for a suitable solution.
If it's woods, and high accuracy is needed, observe it 3x. Notice the shot spread. Use the average function.
Typically, you are going to have it within 0.05' horizontaly, via this means.
Do you have screen shots turned on, in the boundary profile?
Have you yet learned to output a pdf report?
This is (IMHO) the best gps yet.
However, it is still going to have a few hundredths error, and a few more, in bad environments.
Another thing, I've been told that there is a little fuzz, or noise, from all the electronics in the box of the LS. However, over some 20 or 40 seconds, this averages out. This fuzz shows up on as some of the shot spread, on the LS screen.
N
1.) Accuracy is degraded, by a few hundredths. And, sometimes a tenth, and even up to 0.16' The worst I've experienced was 0.14'.
2.) There are only certian times of the day, when real bad places are observable.
So... Certian places may not be suitable for gps. They can yield large error ellipses, and take excessive amounts of time.
Places that come to mind, where this might be true, is between tall buildings, up against large structures, inside buildings. Under external stairs.
There simply is not enough data getting through, for a suitable solution.
If it's woods, and high accuracy is needed, observe it 3x. Notice the shot spread. Use the average function.
Typically, you are going to have it within 0.05' horizontaly, via this means.
Do you have screen shots turned on, in the boundary profile?
Have you yet learned to output a pdf report?
This is (IMHO) the best gps yet.
However, it is still going to have a few hundredths error, and a few more, in bad environments.
Another thing, I've been told that there is a little fuzz, or noise, from all the electronics in the box of the LS. However, over some 20 or 40 seconds, this averages out. This fuzz shows up on as some of the shot spread, on the LS screen.
N
Thanks for the replies guys and sharing your typical process. I live on 8.5 acres and have surveyed every square inch of the place with every piece of equipment I have owned from the early days of Nikon DTM, to Leica 1100, then Sokkia GSR2700ISX, then Leica Viva equipment and now the LS. I have a lot of control to work with and ready checks to everything I shoot. I will say with this, now the 3rd incarnation of GPS, it is the only system I have been able to push to extremes in the woods. I have been getting long durations and multiple sessions as I can set it up and go do other things while it cooks. I have been setting up profiles and do record screenshots.
I have been working with the parameters of the minimum engines to fix with the various confidence levels and various validation settings and whether to reset the engines or not.
This leads to my next question; inside canopy, is it better to accept less data (higher min of fixed engines and higher validation rate) or is it better to accept a lot of data that has lower fixed engines and validation rates?
John
I have been working with the parameters of the minimum engines to fix with the various confidence levels and various validation settings and whether to reset the engines or not.
This leads to my next question; inside canopy, is it better to accept less data (higher min of fixed engines and higher validation rate) or is it better to accept a lot of data that has lower fixed engines and validation rates?
John
In canopy, particularly thick canopy, you may find it difficult to get a high number of fixed engines at one time. I typically require only 2 fixed engines for working in canopy, but I make sure the receiver has at least three minutes between the first fix and the last validation.
More engines will improve reliability, but doesn't guarantee it. What I've noticed is that a bad fix with 6 fixed engines doesn't last very long. After a few seconds, you'll noticed that the number of fixed engines begin to drop. Eventually down to two engines and then float again. Sometimes you will see a fix with 6 fixed engines drop, perhaps down to three, then bump back up to four or more after a few seconds with an agreeable fix. This is a good sign because the new fix came several seconds after the initial fix. These are just "pointers" to a good fix in my mind and not a guaranteed verification. The best verification I've seen is waiting until validate checks at the end. The verification at the front just helps insure you are likely starting on the right course.
More engines will improve reliability, but doesn't guarantee it. What I've noticed is that a bad fix with 6 fixed engines doesn't last very long. After a few seconds, you'll noticed that the number of fixed engines begin to drop. Eventually down to two engines and then float again. Sometimes you will see a fix with 6 fixed engines drop, perhaps down to three, then bump back up to four or more after a few seconds with an agreeable fix. This is a good sign because the new fix came several seconds after the initial fix. These are just "pointers" to a good fix in my mind and not a guaranteed verification. The best verification I've seen is waiting until validate checks at the end. The verification at the front just helps insure you are likely starting on the right course.
Nate The Surveyor
Well-Known Member
First off, John, you are the exception. I absolutely love the fact that you are "Going out there and testing it" against stuff you already know. This is what a RESPONSIBLE and WISE surveyor does.
Second, the answer to your question is that my minimum number is 2 Engines, for my "Going in the bad" profile. I feel it is better to GET more data, with less engines.
HOWEVER...
having said that,
There is a time to set a min of 3 engines. Such as topo in the woods. It does tend to UP the certainty factor. This is not real heavy woods, and you are keeping track of your "Dist to Last" and elev diff.
So, those are the 2 settings I use mostly, is 2 and 3.
Since I don't do alot of topo, mostly 2.
Another place for a min of 3 is when running a centerline of a road, and there are obstructions overgrown. I often shoot 2x as many shots as necessary, to weed out the bad... Maybe John Evers will chime in here... He has experience with the question at hand, relating to a good topo setting.
Hey, ain't it great to have the finest, and funnest GPS in the world?
OOPS... Shawn posted while I was typing....
N
Second, the answer to your question is that my minimum number is 2 Engines, for my "Going in the bad" profile. I feel it is better to GET more data, with less engines.
HOWEVER...
having said that,
There is a time to set a min of 3 engines. Such as topo in the woods. It does tend to UP the certainty factor. This is not real heavy woods, and you are keeping track of your "Dist to Last" and elev diff.
So, those are the 2 settings I use mostly, is 2 and 3.
Since I don't do alot of topo, mostly 2.
Another place for a min of 3 is when running a centerline of a road, and there are obstructions overgrown. I often shoot 2x as many shots as necessary, to weed out the bad... Maybe John Evers will chime in here... He has experience with the question at hand, relating to a good topo setting.
Hey, ain't it great to have the finest, and funnest GPS in the world?
OOPS... Shawn posted while I was typing....
N
I run 1 engine in my boundary and control profile and have great success. I have the minumum duration for verification set to 60 seconds and a minimum duration of 240 seconds for the whole observation. Lots of places I go 1 engine is all your going to get. Once the verification phase has met its requirements J field is going to kick out the epochs that don't agree so it's not a bad thing to allow one engine but you have to be there long enough for the LS to do its thing.
I would say using a one engine fix for a short shot is a horrible mistake. It's like a snake waiting to strike. In my topo profile I have it set to 3 engines. If you aren't there long enough (3-4 minutes or longer) you might be storing a bad fix. I love the distance to last white button. Most looked at button on the screen for me.
I would say using a one engine fix for a short shot is a horrible mistake. It's like a snake waiting to strike. In my topo profile I have it set to 3 engines. If you aren't there long enough (3-4 minutes or longer) you might be storing a bad fix. I love the distance to last white button. Most looked at button on the screen for me.
Darren Clemons
Well-Known Member
All good stuff guys. What it says more than anything is what's so great about the LS - there's no one right way to set it up. It has so many options and with the different profiles you can literally set it up for every scenario and only have to click one button!
I use two minimum engines as well on my boundary profile, three minimum on topo profile, but with what Adam describes one engine could get the job done with enough time separation and validation. That's what's sets the LS apart from other systems. It weeds out the bad data on its own!
I've seen exactly the same as what Shawn describes on those "bad" six engine fixes. We've had maybe three or four (in 20 months time) that have actually went completely through validation and stored what turned out to be a bad shot but at that time I had it set to stop after 120 epochs, which can all be done in only 30+ seconds. These shots most likely won't validate and store if the time is set to 120 seconds or more.
I also agree with Adam on dist to last white box, most looked at button for me too. After most every shot I store in canopy, I immediately hit start again and watch the dist to last. As soon as I see a few epochs in the lead group which gives me 0.04' in the white box I know I'm good to go. Plus, as all have said, by that time I've added more time to my session which again helps in the confidence.
It all comes down to what it takes to make you feel comfortable. If you have control and points you've surveyed many times, then you're way ahead. You know what, where and how accurate your points are so there's no guesswork there.
I use two minimum engines as well on my boundary profile, three minimum on topo profile, but with what Adam describes one engine could get the job done with enough time separation and validation. That's what's sets the LS apart from other systems. It weeds out the bad data on its own!
I've seen exactly the same as what Shawn describes on those "bad" six engine fixes. We've had maybe three or four (in 20 months time) that have actually went completely through validation and stored what turned out to be a bad shot but at that time I had it set to stop after 120 epochs, which can all be done in only 30+ seconds. These shots most likely won't validate and store if the time is set to 120 seconds or more.
I also agree with Adam on dist to last white box, most looked at button for me too. After most every shot I store in canopy, I immediately hit start again and watch the dist to last. As soon as I see a few epochs in the lead group which gives me 0.04' in the white box I know I'm good to go. Plus, as all have said, by that time I've added more time to my session which again helps in the confidence.
It all comes down to what it takes to make you feel comfortable. If you have control and points you've surveyed many times, then you're way ahead. You know what, where and how accurate your points are so there's no guesswork there.
I agree with @Nate The Surveyor regarding testing where you know the answer. I encourage every user to do this repeatedly until they develop the confidence to know that the results they are getting are 100% reliable and accurate. There is no better training and unfortunately too few surveyors act on this advice, anxious to get started on the ROI of their investment. Very commendable, John.
As Nate, @Adam and @Darren Clemons have said, the Distance to Last button is extremely helpful. It provides a sanity check to the last shot when you are doing topographic work in the trees. I also use it just like Darren describes by resetting engines after I've stored a point in a bad environment and looking at the DTL. I can see immediately that it checks with the point I just collected, giving me just a little more confidence.
This is a partnership between man and machine and the more you use it, the more that seemingly inconsequential details will become nuanced indications of success or failure. Are engines fixes slowly increasing or decreasing? Is the scatter plot linear or like a thumb print? Is the vertical scatter like a hockey stick or like a sine wave? What is the separation in time between the first fix and last fix? It's not that you can't be productive with it within a couple of days of taking delivery, but as you gain experience over a couple of years, these things become more apparent.
As Nate, @Adam and @Darren Clemons have said, the Distance to Last button is extremely helpful. It provides a sanity check to the last shot when you are doing topographic work in the trees. I also use it just like Darren describes by resetting engines after I've stored a point in a bad environment and looking at the DTL. I can see immediately that it checks with the point I just collected, giving me just a little more confidence.
This is a partnership between man and machine and the more you use it, the more that seemingly inconsequential details will become nuanced indications of success or failure. Are engines fixes slowly increasing or decreasing? Is the scatter plot linear or like a thumb print? Is the vertical scatter like a hockey stick or like a sine wave? What is the separation in time between the first fix and last fix? It's not that you can't be productive with it within a couple of days of taking delivery, but as you gain experience over a couple of years, these things become more apparent.
Awesome feedback guys, thank you. Now I can see the value of the distance to last button. I have been working towards the "feel" for min. engines, epochs, and validation for these shots and have a few profiles set up to work with. And as all of you have said, getting that "feel" for how the engines are moving in and out of fix-float.
Thanks for the compliments on this type of testing. During the worst part of the recession I took the time to learn all the ins and outs of my previous gps equipment. When things turned we hit the ground running with confidence. Plus you don't need to be in business long to understand the potential problems that await poor decisions made with not understanding the equipment. I've dodged a few bullets in the 20 years I've worked for myself.
What I figured out very quickly is that a very smart man has provided us with a high end scientific instrument. It is up to us as the scientists to set up the parameters to achieve the desired results. What this also means is it takes an experienced and sophisticated surveyor to understand when the numbers are NOT correct. I believe Javad has extended the useful life of all of us as we will be the few who understand the intricacies of this type of data processing. All of my contractor clients are now all doing their own layout with machine control and rovers. All they need are 4 points around the site. But what happens next to the building and along the tree line? Stay tuned!!
Thanks for the compliments on this type of testing. During the worst part of the recession I took the time to learn all the ins and outs of my previous gps equipment. When things turned we hit the ground running with confidence. Plus you don't need to be in business long to understand the potential problems that await poor decisions made with not understanding the equipment. I've dodged a few bullets in the 20 years I've worked for myself.
What I figured out very quickly is that a very smart man has provided us with a high end scientific instrument. It is up to us as the scientists to set up the parameters to achieve the desired results. What this also means is it takes an experienced and sophisticated surveyor to understand when the numbers are NOT correct. I believe Javad has extended the useful life of all of us as we will be the few who understand the intricacies of this type of data processing. All of my contractor clients are now all doing their own layout with machine control and rovers. All they need are 4 points around the site. But what happens next to the building and along the tree line? Stay tuned!!
It is up to us as the scientists to set up the parameters to achieve the desired results. What this also means is it takes an experienced and sophisticated surveyor to understand when the numbers are NOT correct.
I cannot sum it up any better than this. You are dead on. A surveyor who does his home work on the testing grounds, like you describe, will be a top-tier professional.
I cannot sum it up any better than this. You are dead on. A surveyor who does his home work on the testing grounds, like you describe, will be a top-tier professional.
I have been testing the LS/DPOS more in tall, fairly thick canopy and I am starting to see a trend. When I set the parameters of the LS to its max engines, 6 with reset, and max verification, and let the LS collect data for a long time (30 mins +/-) I get an RTK solution that may not have met all the verification parameters, but is a good solution none the less. The DPOS solution however is very usable and the cluster is even better.
I found that when I "forced" the rtk with lower parameters it did not jive with the DPOS. That was bothering me.
John
I found that when I "forced" the rtk with lower parameters it did not jive with the DPOS. That was bothering me.
John