Simple Answer to RTK Uncertainty
> Redundancy.
But the question is how do the RTK users deal with the uncertainties in the positions obtained via RTK methods.
Simple Question for RTK Users/Enthusers - Diagram misleading
Duly noted.
Simple Answer
Simple Answer
>
LOL!
Speaking of Dr. Bock
> The subject of RTK and PPP came up while discussing something else and he has firm confidence that RTK achieves better than a tenth (at 95%) and is a little amused that folks still think that it can't.
Yes, so does it logically follow that everyone with RTK should always expect those results? :-O
Speaking of Dr. Bock
> I'm sure that some testing as I see mentioned in this thread would help answer that.
I doubt very much that Yehuda Bock has even the slightest inkling how RTK is actually used by the employees of land surveyors in everyday work. :>
Speaking of Dr. Bock
> I can agree about fools with tools :>
Well, if he doesn't avoid you when you see him again, why not ask Bock how he would go about estimating the uncertainties in positions determined by RTK methods? On the subject of the optimism of RTK controller estimates of positional uncertainty, ask him for suggestions as to how to most efficiently correct the optimistic uncertainties to reality.
Speaking of Dr. Bock
None of the total stations I've used had any uncertainty read out while taking a shot. I either didn't turn the function on or Kent has a custom model. I wonder if he trusts it! Maybe it's that ppm thing or inches of Hg. I never could figure out what the heck that was all about. Is that the uncertainties? It's saved in the file but not for every shot and there is no covarience numbers, at least in total station files I've used. But I don't have recent model.
Combinations of C-E
We use three, 2 to 3 minute observations on control points and corners, then average those and look at the standard errors that it determines. Two shots only guarantees that you'll get the same wrong answer twice. The third shot, under a different constellation, is the best way to verify the position and is the method we use with network RTK, which we use nearly daily.
In running between these pairs, rather than the standard RTK pairs we used to set (which closed fine), we are finding closures approaching that of running between static pairs, and I have the answer in real time.
You're missing the boat here.
> We use three, 2 to 3 minute observations on control points and corners, then average those and look at the standard errors that it determines. Two shots only guarantees that you'll get the same wrong answer twice. The third shot, under a different constellation, is the best way to verify the position and is the method we use with network RTK, which we use nearly daily.
So, are the three sessions of 2 to 3 minutes duration that you are averaging taken in succession or at different times during the day or over several days?
You aren't using the RTK vectors, but are estimating the uncertainty of the average coordinates from the standard error of three positions? What do you do with that uncertainty estimate other than simply inspect it?
Speaking of Dr. Bock
> None of the total stations I've used had any uncertainty read out while taking a shot.
Well, considering that the uncertainties in distances measured by most modern high-quality total stations are better than +/-2mm + 2ppm, compared to RTK those measurements would be considered to be exact. :>
The rest of us can take the uncertainties of both angles and distances into account when the measurements are run through a least squares adjustment. It's simple since a total station will give quite uniform results that don't vary from hour to hour throughout the workday or from place to place as happens with RTK.
> Combinations of C-E
Options (c) and (e) were:
(c) check coordinates from two or more different occupations and average numbers, assuming that error cannot be greater than differences in coordinates,
(e) export vectors with covariances generated by RTK controller to use in weighted adjustment by software that generates realistic estimates of uncertainties.
Since if one follows method (e), there wouldn't be any point in (c), does that mean that you mostly use method (c), but occasionally resort to (e)? Under what circumstance do you abandon (c) in favor of (e)?
Simple Answer
Never saw that one before, I lost it pretty good in the office here.
> You aren't using the RTK vectors, but are estimating the uncertainty of the average coordinates from the standard error of three positions? What do you do with that uncertainty estimate other than simply inspect it?
Evaluate it after I inspect it to determine what I do next. 🙂
Thanks Lee, I will take another look at the options, especially the "store another". Early on in my GPS experience I used the mean option but it made me nervous because I sat the base on a meaned point the next day and got an error message indicating the point was different than the reported values or some such thing.
> > You aren't using the RTK vectors, but are estimating the uncertainty of the average coordinates from the standard error of three positions? What do you do with that uncertainty estimate other than simply inspect it?
>
>
> Evaluate it after I inspect it to determine what I do next.
Sort of like making it up as you go along? That would be more of a (c), I think.
