> Except that the standard errors of angles and distances measured with a total station in fact can be well estimated from repeat measurements. The only qualification is that the repetitions are made under conditions tending to cancel certain types of errors known to be present in the electro-optical system, such as measuring directions on different parts of the circle or choosing multiple intervals for distances with some geometric contraints such as that all of the intervals lie on the same line.
>
You completely missed my point. If I'm in the field and measure two angles with a 3 second gun that fall within one second of one another, I'm not going to assume that those angles are accurate to one second. If in the field, I measure a distance with a 3mm+2ppm EDM twice that agrees within 0.001', I'm not going to assume that the distance is accurate to 0.001'. If in the field, I measure a position with RTK that is repeated to a millimeter, I'm not going to assume that the positions are accurate to the millimeter. This is what your option C afforded and I reject it.
> Wanting to consider RTK a sort of GNSS "total station" is not a very good analogy, even if the GNSS salesfolk may still be using it.
>
I'm not sure how you arrived at this analogy. I certainly did not make it. I made an analogy to limited repeat measurements not being useful for determination of accuracy.
Kent I believe you are correct about the unit variance. The dx dy dz values are contained in the QC1 record; when you set up an RTK survey with Trimble you get the option to store one or both - Trimble states (obviously) that you need to store both if you wish to perform a LS adjustment on the RTK vectors.
Kevin -
Trimble Access can also store at a specified time interval; the only problem with that is that it's intended to be used in a dynamic way and therefore only stores one epoch of data. I'm not aware of any way to change that.
Speaking as someone with 20 years' experience using RTK, I am 100% convinced, as I posted on another thread, that it is mathematically impossible for two fully independent RTK observations to give you the same wrong answer within the manufacturer's specifications. If I have redundant observations on a point over a period of time and the standard deviations and coordinate qualities reported are at an acceptable level I have no reason whatsoever to believe that the uncertainties on the point are anything more or less than what I'm seeing. The definition of "Acceptable level" might change based on conditions and what it is I'm trying to accomplish; obviously best practices must be used in order to achieve the highest precision. All a Least Squares adjustment is going to accomplish is to offer further confirmation of what I already know.
I know with absolute certainty that I can achieve certain levels of precision with RTK using proper care and techniques. If that precision isn't good enough for what we're doing, then RTK isn't the right tool for the job. Part of being a professional is understanding your tools and knowing when to use what to achieve results that are satisfactory in the most efficient manner possible. There aren't too many things that I'm going to spend twice as much time on (or more) in the hope of improving the results at sub-cm levels.
All of the theoretical discussions about uncertainties and covariance matrices are great, but at some point it boils down to taking an absolutely fabulous tool outside and making money with it.
My $0.02
:good:
> You completely missed my point. If I'm in the field and measure two angles with a 3 second gun that fall within one second of one another, I'm not going to assume that those angles are accurate to one second.
No, you'd use the standard test procedure to determine the standard error of a direction observed with that instrument. That test procedure involves repeated measurements of directions. In other words, repetition used properly is an excellent tool for evaluating uncertainty.
In the case of RTK, the testing is complicated by all of the external factors that effect RTK solutions. Meaningful repetition takes care to sample under various states of those external factors.
In TBC there is not an option to export this data when setting up and export format.
Is there a way to extract this data via TBC or did you have to parse a XML file?
:good:
Shawn - RTK Surveying - Small Example
:good:
> If I have redundant observations on a point over a period of time and the standard deviations and coordinate qualities reported are at an acceptable level I have no reason whatsoever to believe that the uncertainties on the point are anything more or less than what I'm seeing. The definition of "Acceptable level" might change based on conditions and what it is I'm trying to accomplish; obviously best practices must be used in order to achieve the highest precision.
Basically, what you're applying is a non-quantitative measure to uncertainty, asking whether things are just the ordinary errors or not. And if they're larger than ordinary, you're asking "does it matter"?
That is pretty much a (b) or (c) option in the heirarchy of ways of treating uncertainty.
> All a Least Squares adjustment is going to accomplish is to offer further confirmation of what I already know.
Actually, that wouldn't be true since a least squares adjustment can provide a method for quantifying and testing uncertainty in a way that doesn't have the personal biases you describe in your own methods.
Shawn - RTK Surveying - Small Example
Thank you for that, Gavin. I agree entirely.
I had an email conversation with a gentleman (that I do not believe participates on this board) suggesting that LDP's were a wasted effort as the accuracy of GNSS wasn't worth parsing hundredths. He had considered all of the error sources (from centering, to plumbing, to baseline errors) and felt 3cm was about as good as one could expect between to rover points (not directly tied together). I told him that I was sorry that this was his experience. In open skies, I routinely find that two RTK rover points with 3 minute occupations (possibly overkill) can provide a baseline that is accurate to within about 5mm when measured with my total station. I do this almost every day.
Adding multipath to the testing is a little trickier because it is near impossible to quantify multipath (at least in my primitive mind). So experiments such as you recommend (head to head - no pun intended) is a better way to test multipath mitigation.
On your CBL tests, do you have the base at one end and the rover at the various distances, or is the base remote (or RTN) and you are comparing inverse distances between positions?
Shawn - RTK Surveying - Small Example
It's amazing to me that the hard and fast "rules" we apply to GNSS (I struggle not to say GPS anymore) change so rapidly. I haven't been using RTK for long, perhaps 7-8 years of experience with different manufacturers. In my early experience, I saw most receivers (base/rover) had a functional precision of about 1.5cm horizontally. Then I reviewed gear from Javad, and then Altus, and Trimble that were getting very sub-centimeter results. Suddenly post-processing didn't seem so attractive - marginally better results in 10 times the time. I still use post-processing when it's appropriate.
PPP is some amazing stuff. It'll change everything... again. 🙂
Shawn - RTK Surveying - Small Example
I'll gladly buy the first round.
Gavin & Shawn
I think that you guys put your respective Latitude & longitudes into Star*Net (with the right "reality" fudge factors), you will end up in Evanston Wyoming (or there about), in which case I'll buy ALL of the beer!
Loyal
Shawn & Loyal
I'm with Gavin on that. How's the weather in Wyoming this time of year?
Shawn
Sunny & mid-high 70s today...not too sweet for the next few days.
Shawn
Nice wrap up G & S & L. Wyoming huh? We have been thinking about a trip out west...
Loyal-Brad-Shawn - P.I.N.G.S.
I can't guarantee things won't get hysterical, Gavin.
Shawn
:good: