Activity Feed › Discussion Forums › GNSS & Geodesy › Testing Rover Pole Centering Accuracy
-
Testing Rover Pole Centering Accuracy
Posted by Kent McMillan on July 31, 2016 at 3:10 pmI’m going to guess that pretty much all of the folks who survey boundary control via RTK and PPK methods use a rover mounted on a pole. Setting aside the matter of corrections of centering via tilt sensors and the environments in which they work or don’t work, the basic question is how well one can center a GPS antenna on a rover pole at various heights.
Has anyone ever tested the centering accuracy of their rover pole to derive the standard errors of centering, either DRMS (standard error of radial error) or standard error of the orthogonal components of the total centering error? If so, what values did you determine to apply?
conrad replied 8 years ago 12 Members · 33 Replies -
33 Replies
-
Since I don’t survey pianos, I just rotate the pole between shots and call it good.
-
I don’t trust the tilt sensors to be useful for anything other than rough work, so I ignore them. I mount the receiver on the same pole I use with a prism, and use the pole’s 10-minute bubble for centering. There’s a small amount of play introduced by the Leica quick-release mount, but it’s pretty negligible.
-
That’s a question I ask myself every time when I see folks using a tall bamboo stick to put the GPS above the tree canopy. RTK say it’s good to a cm, but that’s not necessarily applicable at the bottom of that 10-m stick…
I can only think of doing successive reoccupations and calculating the standard deviation, assuming it’s only random errors. Perhaps even using a triangle template on the ground with known side lengths to impose more randomization.
Chapter 4 in Ogundare (2015) “Precision Surveying” gives this:
<“>http://onlinelibrary.wiley.com/book/10.1002/9781119147770>
4.5.4 Instrument and Target Centering Errors
The string plumb bobs in windless weather condition will give a [horizontal] centering error of 1mm per height of instrument in meters; it will be quite worse in windy weather.
(…)
It should also be understood that themiscentering error is systematic for an individual setup and it occurs on every pointing. Target miscentering error cannot be reduced in size by multiple pointings, but the error will appear random over multiple setups on a point. In this case, the effect of centering will be randomized and reduced by the square root of the number of independent recentering done in the multiple setups while measuring the observable. For example, if an instrument is recentered four times over a point while measuring an angle, the centering error in the angle will be reduced by 2.-FGN.
-
Kent McMillan, post: 383480, member: 3 wrote: I’m going to guess that pretty much all of the folks who survey boundary control via RTK and PPK methods use a rover mounted on a pole. Setting aside the matter of corrections of centering via tilt sensors and the environments in which they work or don’t work, the basic question is how well one can center a GPS antenna on a rover pole at various heights.
Has anyone ever tested the centering accuracy of their rover pole to derive the standard errors of centering, either DRMS (standard error of radial error) or standard error of the orthogonal components of the total centering error? If so, what values did you determine to apply?
Hello Kent,
I’m sure others have better methods but to test the R10s level bubble against the bubble on the rod and the rod itself I set up our robot in the office and sight bottom to top revolving 360. I’d be interested in what you find out. Thanks, Paul -
rlshound, post: 383511, member: 6800 wrote: I’m sure others have better methods but to test the R10s level bubble against the bubble on the rod and the rod itself I set up our robot in the office and sight bottom to top revolving 360. I’d be interested in what you find out.
I’d think that the test procedure would depend upon how the user was typically plumbing the rod, i.e. whether he was consistently in one orientation such as standing right behind the bubble on the axis containing the level vial and the center of the pole, or whether he was standing in some random relationship alongside the vial differing from setup to setup.
The way to test the bubble that would seem generally most revealing would be in two axes, one to determine random errors of centering along the axis from bubble to center of pole and one along the axis transverse to that. My hypothesis would be that the observer can plumb better side to side than along the axis from bubble to pole where parallax enters the picture in viewing the bubble.
This is something I’ve never tested since I use prism pole tripods, but it would need to be taken into account in any freehand RTK work.
-
Kent McMillan, post: 383513, member: 3 wrote: it would need to be taken into account in any freehand RTK work
If it’s truly freehand — no bracing at all — then I don’t think trying to quantify it would be worth the effort.
-
It’s not just how good is the bubble but how straight is the pole from tip to target/ antenna.
Given accuracies or repeatability of RTK observations then it’d be a pretty rough pole to exceed that dimension.I use a hand pole bubble and alligator clipped tripod for anything I consider serious.
I’ve also taken shots at each “quadrant”. Rotating the pole 90å¡ between successive readings.When GPS matches EDM then I guess it’ll be critical, but all this also applies to EDM pole measurements.
-
Jim Frame, post: 383515, member: 10 wrote: If it’s truly freehand — no bracing at all — then I don’t think trying to quantify it would be worth the effort.
You could quantify the standard error of centering of an antenna on a freehand rover pole, though. If you think that only bipods are used, of course that would be the more revealing value to test.
It would be a bit complex to model since you’d have the quasi-random error of bubble maladjustment with the random errors of bubble reading added to it. Naturally, the bubble maladjustment largely cancels when prism pole tripods are used, but would be present with a bipod.
-
Richard, post: 383516, member: 833 wrote: It’s not just how good is the bubble but how straight is the pole from tip to target/ antenna.
Given accuracies or repeatability of RTK observations then it’d be a pretty rough pole to exceed that dimension.I use a hand pole bubble and alligator clipped tripod for anything I consider serious.
I’ve also taken shots at each “quadrant”. Rotating the pole 90å¡ between successive readings.When GPS matches EDM then I guess it’ll be critical, but all this also applies to EDM pole measurements.
The question is whether the RTK positioning errors are sufficiently large to swamp the rod centering errors. As you say, if the RTK positioning errors are +/-1cm DRMS, then random errors of rod centering would need to be more than about 6mm DRMS (about 0.02 ft. DRMS) to make about a 20% increase in the resulting errors of positioning.
-
Kent McMillan, post: 383513, member: 3 wrote: I’d think that the test procedure would depend upon how the user was typically plumbing the rod, i.e. whether he was consistently in one orientation such as standing right behind the bubble on the axis containing the level vial and the center of the pole, or whether he was standing in some random relationship alongside the vial differing from setup to setup.
The way to test the bubble that would seem generally most revealing would be in two axes, one to determine random errors of centering along the axis from bubble to center of pole and one along the axis transverse to that. My hypothesis would be that the observer can plumb better side to side than along the axis from bubble to pole where parallax enters the picture in viewing the bubble.
This is something I’ve never tested since I use prism pole tripods, but it would need to be taken into account in any freehand RTK work.
Sorry, let me clarify, left out this part…the pole is braced within a tripod….looking along both sides top to bottom and comparing to center…
-
rlshound, post: 383541, member: 6800 wrote: Sorry, let me clarify, left out this part…the pole is braced within a tripod….looking along both sides top to bottom and comparing to center…
If the pole is always set up within a tripod so that it can be rotated to verify plumbing, that’s a different situation than if the plumbing is accomplished by observing the bubble in just one position. If in just one position, I think I’d want to determine the standard errors of plumbing in two different directions: axis from bubble to pole and side-to-side axis perpendicular to first.
It would involve doing as you suggest, i.e. measuring the angles from the tip to the top of the pole at a known distance from the pole and computing the deviations from plumb of successive setups (at least ten) from the angles and distance. Then, the pole would get rotated 90å¡ and the centering errors of another series of setups would be measured by the same methods.
From the series, the standard errors could be calculated. It would also be necessary to see if the centering deviations show a systematic error.
-
Kent McMillan, post: 383544, member: 3 wrote: measuring the angles from the tip to the top of the pole at a known distance from the pole and computing the deviations from plumb of successive setups (at least ten) from the angles and distance
I’ll reiterate my earlier comment with a question: why spend time determining the sub-millimeter centering error of a device used to support a centimeter-level instrument? In my opinion, if you have a straight rod with a 10-minute bubble and the bubble checks visually when rotated against a plumb fixture, you’ve done all you need to do to support an RTK antenna.
-
I’ll second that Jim.
It boils down to knowing your gear, it’s capabilities, weaknesses and being confident in your methodology.
That’s something often lacking today.
Where people are given basic training to read the script that is written to prevent idiotic decisions made by those that have little understanding of the ‘thing’ in their hands /control.With regards to Kent’s comment I’d suggest GPS measurement uncertainty would “swamp” pole errors. Unless it was a hopelessly calibrated bubble which surveyors should check frequently.
-
-
Jim Frame, post: 383546, member: 10 wrote: I’ll reiterate my earlier comment with a question: why spend time determining the sub-millimeter centering error of a device used to support a centimeter-level instrument? In my opinion, if you have a straight rod with a 10-minute bubble and the bubble checks visually when rotated against a plumb fixture, you’ve done all you need to do to support an RTK antenna.
The answer is that almost certainly the centering accuracy of a rover antenna on a pole, both freehand and in bipod, is completely unlikely to be sub-millimeter. For a freehand pole, wouldn’t you expect that centering errors could approach 1cm, depending? So, if you ignore that error component and dwell on the accuracy of the vector to the antenna, what does that accomplish?
-
Richard, post: 383550, member: 833 wrote: With regards to Kent’s comment I’d suggest GPS measurement uncertainty would “swamp” pole errors. Unless it was a hopelessly calibrated bubble which surveyors should check frequently.
Isn’t the point, though, that if one knows that the pole centering errors do not exceed a certain value, then their effect on positioning can be also reasonably estimated. However, if one does not know what the pole centering errors actually are (particularly with a freehand pole with a systematic error), then it’s just mostly guesswork?
-
Kent McMillan, post: 383574, member: 3 wrote: The answer is that almost certainly the centering accuracy of a rover antenna on a pole, both freehand and in bipod, is completely unlikely to be sub-millimeter.
Not the way I use a bipod. Which reminds me: I wish Seco would make a bipod with micro-adjusters like the ones they use on their heavy-duty tripod. The backlash in the thumb-button mechanism makes it annoyingly difficult to fine-tune the setup.
-
Jim Frame, post: 383578, member: 10 wrote: Not the way I use a bipod. Which reminds me: I wish Seco would make a bipod with micro-adjusters like the ones they use on their heavy-duty tripod. The backlash in the thumb-button mechanism makes it annoyingly difficult to fine-tune the setup.
So, does that mean that you’ve tested the centering accuracy of your rover pole in bipod and found that it has a standard error below 1mm? It seems unlikely that accuracy would be possible without fine tuning the setup by reversing the bubble position (which cannot be done, insofar as I know, in a bipod).
-
Kent McMillan, post: 383580, member: 3 wrote: does that mean that you’ve tested the centering accuracy of your rover pole in bipod and found that it has a standard error below 1mm?
I’ve tested the pole in a tripod. The bubble accuracy doesn’t change when I move it to a bipod.
Log in to reply.