Activity Feed › Discussion Forums › GNSS & Geodesy › How Long Between GPS Control/Boundary Corner Occupations?
How Long Between GPS Control/Boundary Corner Occupations?
OleManRiver replied 1 year ago 23 Members · 78 Replies
I think you guys are over thinking my request (but thanks!).
I want my guys to quit giving me two 5 minute sessions 2 minutes apart.
I am curious how long it takes to get a different “satellite configuration” in the sky. I mean enough different satellites and/or different locations from one occupation (with a few 120 sec sessions) to the next occupation (with a few 120 sec sessions) to provide some “meaningful redundancy”.
I think I’ll just go watch a skyplot in the DC and watch the time. Maybe I’ll wait for a warm Friday afternoon in the back yard with a cooler?
If you have a safe spot at the office or at home, just set the unit(s) up and set the data collector to record a X second observation every Y minutes for the day. You’ll have a whole bunch of observations to work with and can see if there is some time spacing that you can define as being “better” to wait. Also, you’ll be able to see if there is really any improvement at all by requiring some time between observations.
You could record a static session as well and look in your processing software for how the satellites shift within quick glance at the data instead of having to hang out watching the collector screen.
That way, you can still have the Friday afternoon cooler, but also spend time grilling some dinner up as well instead of checking a screen every few minutes.
@scott-bordenet
From one of the several papers I posted upthread:
“Double occupations are well established, best practice for many surveying applications, including GNSS observations. They can detect blunders, like observing on the wrong mark, poor centring or a wrong instrument height. For GNSS, double occupations are also useful to detect the effects caused by incorrect ambiguity resolution or bad multipath conditions…
…As mentioned earlier, the averaging technique can still produce a result sig-nificantly offset from the actual position. For high-accuracy applications, double occupations can therefore improve the precision, accuracy and reliability of GNSS positioning…
…But how long should a user wait until re-observing a mark? Our analysis of the same data indicates that two occupations can be assumed sufficiently independent from each other if they are taken 10-30 minutes apart. Waiting any longer to re-observe is generally not likely to improve positioning results
any further. This applies to both RTK and NRTK.”“…people will come to love their oppression, to adore the technologies that undo their capacities to think.” -Neil PostmanNGS has a sample data set that shows observation for a long duration GPS only I believe. But you can see the coordinates from each hour if I remember correctly. I played with the numbers one night for fun. Any two shots 4 hours apart was closer to truth accuracy to the datum. I remember the spread between shots 4 hours apart was greatest. 2 hours closer relative but not accurate to the truth the datum. 1 hour man shots were very close but further from the truth accuracy the datum. Land surveys are more about precision than accuracy to the datum. So all the rotate the pole 90 or 180 is checking the plumpness of rod and gets you better relative precision. Which is perfectly acceptable as the truth is the property monuments and deed. If you want accuracy to datum then 4 hours is best no matter what gnss rtk base and rover network rtk etc. its a function of orbits. I think 2 hours gives plenty the constellation has changed enough that the multi-path possibilities are not repeatable. I have taking shots for 1 hour straight every 30 seconds and come back 4 hours later to repeat in a multi-path environment and they don’t match. They were relative for the hour nice and repeatable. But not between the gap of time. When i am in canopy now i might take a few different shots back to back. Then come back a couple hours later and do the same. I always compare the spread by the largest time gaps. Rovers paper is probably a good study but 10 to 30 minutes is still mostly in the precision range not complete accuracy to a datum. Now that knowledge is for in in gps glonass only as I understand how the orbits and EOP chandler cycle and other things that go into the global reference frame which nad83 igs all use to get to any local datum. Like nad83. Wgs84 is translated to other datums or has historically been that way. As al these new constellations are all trying to work together until they don’t. Like war. Conflicts. I prefer to ignore glonass at the moment just my thinking. Glonass is not even close to the other 3 period. Glonass caused more harm than good to be honest. Take and simply download a 12 hour session from a nearby cors station from ngs just highlite 4 or 5 sv numbers make sure to pick them out of different prn slots and go through until they are less than 2 time wise then choose 4 or 5 more do the same. You will see exactly how long it takes for the constellation to truly change. You want different birds to make constellation change so the recievers are forced to use different ones as each satellite has its own personality if i can say that. They all even the ones built exactly the same way act different. The clocks on the satellites themselves to l1 and l2 bands. Same signal same information but they all have quirks. Everyone here that ran a total station a long time and had a loaner of same exact model saw this. Just something a little different about it. Edm a little slower or little faster tangent screws or the feel when turning angles etc. and remember that svn is sat specific but prn is not so svn 1 can be prn 1 for a while then prn 2 etc. so learn the hick-ups of each bird. Become 1 with the birds of space lol. Trust me every night at a certain time i use to wake up because one bird always had issues at a certain time. I had to watch it for a period of time until it got back to what its supposed to do. A lot of what i say is over kill what we do i know this. We are not doing absolute positioning yet. But A lot of what i thought i knew about gps from surveying was helpful but only one side of the coin. When i started as an analyst monitoring I learned and started looking at it much differently because i saw another side and then getting more into other aspects i saw more and i still don’t know a tenth of all of it. Its fun learning though.
@jitterboogie not far off at times for sure. They play a lot of games as well. Change names and signal messages. A lot of craziness at times. It would not surprise me to see manufacturers not truly use them in a solution but maybe just for show. A particular manufacturer did just that to help move equipment for a while. Once people caught on those who actually paid attention raised a stink about it. Unlike Galileo and bedue and gps they all can theoretically stand on there own like gps so one can use only one of them if equipment is set to do it and get survey grade positions globally glonass is not that capable all over for that purpose it does have a purpose though they might try and catch up again they have been on the right track but lack of funds set them back they take one step forward and two back historically. It will be interesting to watch the Galileo and bedue in the long game bedue is pretty darn solid and has some tremendous technology But time will tell
One thing is when you cannot based on the job and budget wait a while between observations. Is to lower or raise the rover rod by a minimum of .98 ft. I like those rods that are not fixed two piece for this. That gets you some redundancy as you are getting a different wave length signal to solve. Not perfect but we all know sometimes we just have jobs we need to get in and out. You can also when applicable just set up a total station or robot and measure between or cross tie between a sample. I had some common corners that came in on a recent job that were about a foot apart. I always make some direct measurements with a 25’ tape between those and a few that were 10 and 3 ish. Easy to do even solo as a sanity check. I have on jobs also when in a time crunch set robot up and done some resection and incorporate pins and control ties front lines to back or side to side. The biggest thing with gnss rtk especially is Redundancy redundancy redundancy. And you can achieve that with more rtk shots or using a different tool.
Land surveys are more about precision than accuracy to the datum. So all the rotate the pole 90 or 180 is checking the plumpness of rod and gets you better relative precision. Which is perfectly acceptable as the truth is the property monuments and deed.
You’re typing these on your phone, right?
I do a 1 minute shot, rotate rod 180 and reinitialize the satellites, another 1 minute shot and average. I haven’t been tracking this super hard but it seems like about every dozen or so times I do this I get a bad shot where the vertical on one of these shots is .10-.12 off from the other. When that happens I keep taking shots until I get 2 shots in a row of close measurements.
It’s pretty rare that I don’t have to go back to a site at least one more time, so I’ll always do check shots on subsequent visits. If I’m getting similar measurements on following days then I’m feeling really comfortable.
I keep taking shots until I get 2 shots in a row of close measurements.
Two in a row isn’t a good criterion. Getting a majority that agree closely, and not sequential, is better.
.@sergeant-schultz yes sir i am. I go from surveying to farm and the phone is what is with me. Degnabit I need to buy a personal laptop and mount it in the truck so i stop being dumb and corrected by what this stupid i phone thinks I want to say. Lol. .
@mightymoe yes and that might be the case we have a situation now that has good geometry often. But that still doesn’t change the fact that as satellites orbit and new ones come in they are telling us where when they are nor has enough time passed to consider the chandler cycle or other important implications that go into satellites for truth. This is all fine for the precision but if we want absolute truth accuracy we have to consider those things also. Now i agree not probable necessarily needed in a land surveying situation. That I agree. Its not like the satellites are getting realtime adjustments. What happens at the user end when a delta v occurs on any given satellite and it is in your solution? Not all are done with notice. Probably with all constellations the receiver can cast it out. Or reject it. But historically gps could not at user end. These are not the norm but the 1% maybe range. Those anomalies we saw at a different time. What happens and is working in the month of say January may not in june or December as an example. Prn slots have different sv in them. It is one reason any academic or scientific paper published about gps that doesn’t have a full years set of data points or has not ran the test say quarterly and have a full cycle of data 12 to 24hrs. Over several days and different times of the year is not a guarantee yearly for us. We need to keep that in mind. The constellation today at any given time is not the same several months from now. Now all of that is overkill for surveying most of the time. Several studies suggest that up to a hour of time difference increases the results only 15 to 25%. If memory serves me. That also could be gps only it has been a while since I was in that realm of data at my fingertips daily. What I do know is that even though it was done with gps and glonass only the rtk guidelines by ngs with rt class 1 2 and 3 wasn’t just a guess but tested so it is a pretty good way to understand and have something to stand on if needed of how to achieve accuracy at different levels. Again accuracy to the realization of the datum. One thing that blows me away is it now takes 12 to 19 days to get a precise orbit from igs. Ngs use to turn it around in 7 days for conus. I think you can get one quicker from NGA on gps only but thats a little different animal than we use. At least when i was there we did. I think with practicality and from a production standpoint a hour is great most of the time if the base is moved. So you get a little redundancy for comparison. I have done the back to back with resetting satellites tracking on jobs. Sometimes we just don’t have it in the budget to wait. I try and tell my crews first thing is set a few points i don’t care if its a boundary corner or control around site then do the rest of the work mapping topo other corners and try and mov that base to one of them to hit a second time if possible. That usually is workable and often is more than a hour anyway. Set a point in the middle of site clear and base on it radially tie everything like a wagon wheel. Then set up on one on outer edge and re measure. If its vrs i saw extremely great results after 4 hours to knowns gps glonass. Not today’s full constellation. 2hrs were good as well but 4 hours were almost published values after an adjustment. Or even weighted average. 1 hour was probably good enough for what we do. But again it was on a network i operated so I understood it New its quirks and could personally check each station at the time of making those observations. Thats the biggest con to network rtk is you don’t know if a hickup occurs at any station around you in realtime.
If the constellation’s orbit models are all basically modeled the same, i.g. 24 satellites orbiting the earth twice in 24 hours, thirty minutes should give you 15° of ascension/descension.
Then there’s open skies and a good continuous weather pattern.
Maybe knowing the source base antenna and what constellations it bases it’s broadcast corrections on.
GS’s *apples, oranges, and bananas* analogy, where you don’t want just one R satellite bracing your southeast horizon.
Trimble’s Access lists your VCV components in the raw file. Sometimes you can look at that and see where a component has just gone to H@#$, and why your inverses are adding a few extra hundredths.
Sitting here watching my daughters riding lessons I pulled up the Trimble planning tool. It has a drag bar so you can go by time. And see where sats are coming in going out. I guess after thinking and reading here i should clarify. When i State constellation change its not that sats have just moved they are always moving. A full constellation change for me is all new svs that is how i refer to constellation change. And that makes a difference in my old world. I think here its more of a change in geometry. Which is fine and makes since for the 30 minute type statements here. So poor communication on my part sorry. When i state that new constellation now maybe what i was saying makes more sense. A complete new group of satellites for a measurement. Tis the world i lived in. The way those above me defined it for our purposes. Sorry about that. I see now here most simply mean enough change in geometry.
Most surveyors and their crews around the area I now work in seem to use a single base location, take a 1 to 3 minute (maybe) shot per monument, break and reacquire signal lock and take a 2nd 1 epoch to 1 minute shot and compare it to the 1st.
When I ask them them when they did most of their surveys with a total station if they ran open control or closed traverses, they either say “closed” or “what’s a traverse”, depending on their age and how long they’ve been surveying. If they say closed, I tell them that their method with GPS is about like running an open traverse, hitting the distance button a 2nd time on their observations, reasoning that’s their redundancy so no need to close a traverse.
I always seem to get a blank stare back. When speaking with one local survey business owner a few months back about redundant measurements, he told me that being in the private sector, he can’t justify “surveying it twice.” Huh?
Huh?
I hope everyone has a great day; I know I will!@scott-bordenet
With 4 constellations (2) 5 min sessions will be fine with base/rover with the baselines you are using.
When using 2 constellations I have seen a bad fix repeat itself in 240 seconds maybe 2 times. So even at (2) 5 min observations you would be able to tell that one or the other was bad.
Just my 0.02
With 4 constellations (2) 5 min sessions will be fine with base/rover with the baselines you are using.
I generally advise our crews of the same, with the spacing being a minimum of 30-45 minutes apart in sidereal time. We also instruct our crews to use a RTK-logging survey style so we can post-process if necessary.
We actually had an aerial in the concrete jungle recently where we discarded all RTK observations and used post-processed vectors from CORS instead- party chief placed his base ~150′ north of a 10+ story building, and couldn’t get his “round 1” and “round 2” observations for a large percentage of his targets to agree within a tenth or two. Post-processed vectors cleared up the issue beautifully.
EDIT: I also try to have our crews set a secondary base. If RTK-vectors are good, we can also include post-processed vectors from the “other base” to help tighten things up further or troubleshoot if necessary.
EDIT 2: Generally speaking a 5 minute observation is arguably overkill with RTK. To me the safety blanket of slightly longer observations and the use of RTK+logging allowing me to have complete confidence in our data warrants that overkill. However for surveying in the woods… Your mileage may vary.
For a regular boundary survey, using base/rover, I take five, 30 epoch observations, pausing 30 seconds and ‘dumping’ (really just hitting the reset button) between each. I review the session and eliminate any clear outlier. If there are two obvious outliers, I’ll kick both and take another 3 observations with the same criteria. I generally like to see a standard deviation of under two hundredths on an average with at least 4 separate observations.
Normally the whole observation takes about 5 minutes of occupation. I use Carlson SurvPC; previously I did all of this manually, but in the new version 7 release, that entire session (measure, wait, reset, repeat 5x) can be automated.
The exception would be an ALTA or other ‘high-stakes’ boundary survey. In that case, I will repeat the same pattern, except generally I will make a second lap, this time observing using VRS corrections instead of my local base. If VRS isn’t available then I will try to use a second base setup. The idea being to capture multiple baselines to each corner so that I can better prove the result, but a time delay is a fortunate byproduct.
In my spot checking of this procedure, mostly on sites that I have come back to days later for one reason or another, I have been satisfied with the repeatability of the results. I have an additional spot check built in on every job, a check nail set in the morning and observed as a boundary corner. I generally find my shot at the end of the day to check under .03, often under .02.
Something that puts my mind at ease: Even in ideal conditions, a GNSS survey will not have the repeatability and provability of a well run conventional traverse, but for most boundary surveys, that’s okay. With quality equipment, plumb poles, and good procedures, it can be more than good enough. Boundary surveys are not an accuracy contest against other surveyors, it’s a simple binary question: Does your work meet or exceed the minimum standards for your jurisdiction? I have confidence with my procedures the answer is yes.
A final thought: what’s the next guy going to do? All else being equal, what procedure is he going to apply that will show a result he can prove to be “better” than mine?
The ODOT linked study that @rover83 posted is the most recent (2021) extensive paper of such things that I am aware of.
On Page 48 you will find the mythical two hours mentioned: “When analyzing the time between repeat independent observations the results indicate that a truly independent observation is achieved when an interval of 2 hours between repeat observations is used. Further research is necessary to evaluate the influence of intervals less than two hours between independent repeat observations.”
I have been using the two hour rule on a different day for 33 years now.
SHG
@shelby-h-griggs-pls two hours I believe is a sweet spot for what we need to achieve for sure. If that second observation does not fall within a specific range of the first you would need to wait again or use another method to decide which one was bad. It will get better once all gnss is fully functional and the kinks are worked out. There are still some timing issues which will need to be resolved but if no major conflicts happen I believe it will not be long. The other will be a few years as the other systems get enough datapoints to model the clock errors and drifts on the satellites themselves. Not enough time yet for them to do that. Which is why we are not using all signals that are available now but what the manufacturers can use safely. I noticed on B and gal that on a couple svs only 1 band was actually in the equation on one day same svs next day nothing being used and it was in the same slot. So as good as we see it now we have not even seen exactly how good it will be. Clocks on satellites are like kids each one has its own personality. Then you have the clocks at the monitor stations. This is where gps shines. IGS has more data more stations but more is not always better. I believe one day we will be solving in real-time with no corrections what we are achieving today with base and rover or network rtk solutions. Very fun times.
Log in to reply.