Detail Surveying and Creating GNSS Control
Hey everyone, I have a few questions about GNSS RTK surveying that I was hoping someone could help me with. Just as a disclaimer: I took a few surveying courses back in college, mostly focused on total station work and adjustment calculations, but I haven’t done much actual survey/fieldwork since then. Now I’m back in the field using a Trimble base and rover setup, and I wanted to double-check if my process for setting control and collecting observations is correct:
I start by placing a nail in the ground, setting up my base over it, and running a static observation for about an hour to get a <10cm absolute position. Then I convert those global coordinates and ellipsoidal heights to UTM and CGVD2013 to use as my plane coordinate system. Next, I take my rover, set it up over another nail, and establish a control/check shot by taking 5-minute observations about 20 minutes apart and averaging them.
Each day, when I’m collecting observations, I set up my base over my first control point and do a check shot on the second point at both the start and end of the day to make sure there are no blunders. For the rest of my observations, I typically take two 5-second shots and average them.
Here’s one thing I’m curious about: what’s my expected relative accuracy when following this process? Should I just rely on the R8’s manufacturer specs (10mm + 1ppm for horizontal and 20mm + 1ppm for vertical), factoring in any leveling error from the rover? So, if I repeat observations on the same features over different days, would 1-2cm horizontal and 3cm vertical accuracy be a reasonable expectation?
Also, I remember using total stations to take redundant observations and running least squares adjustments to reduce variance. Is there a similar process with GNSS? Is there any benefit to moving the base and generating redundant baselines (assuming that's even the right approach)?
Any input would be greatly appreciated! I hope my process is mostly correct, and that I’m not totally messing it up. Thanks!
How are you getting the "<10 cm absolute position"?
If you're not processing the static, there isn't much reason to spend an hour on the point. Might as well simply hit the store on the DC and start surveying. You can always collect data on the Base as you do the rover work for later processing. You don't really need to "convert" any raw data to UTM and CGVD2013. Simply set up the DC file with those parameters and when you store the point it will already be in those projections.
If you spend an hour, or ten in the field the base will be collecting data and can be positioned geographically very closely to world and local coordinates after the first survey day by processing the static file that was collecting.
In TBC (if you're using it) reposition the base point after the static processing, press re-calculate and all the collected points (except stake-out points) will be repositioned.
Be careful to correct all the data before deleting the first day in the DC file,
The accuracy specs for the R-82's I use are generally exceeded during my surveys. If you're using R8 units with only GPS I don't have a feel for those.
It's always a good idea to have check-in points as you start and finish the day, it proves your base isn't moving. Check-in points can be done many ways, previous located monuments are also a great way to check for movement. Sometimes we will have a dozen points clustered around an easy to get to monument.
There is a lot going on in that question, but perhaps I can attempt to summarize the questions and we if these are indeed your questions, perhaps get a better set of answers for you?
- What are correct procedures for establishing initial control using static GNSS observations?
- What are the correct procedures for an RTK survey, including checks to existing control, etc?
- What is the appropriate use of a total station in conjunction with static GNSS for initial control?
- How would you use a total station to provide checks for an RTK survey?
Take a look and see if this is what you are asking.
Relevant information that would be needed:
- What software are you using to post-process your observations and work?
- Does that software include tools for balancing a network of observations?
- What is your level of expertise with the software?
- What is the purpose of your survey? (Boundary, etc?)
- Does your project need to be on a certain datum?
And, if I can gently suggest it, all of the topics are covered in past posts. I would suggest searching them first, and then ask questions for clarification or for your specific needs. This is probably the best approach because you will actually have access to more information and guidance than whomever sees this currently.
Some good questions, but this is a bit of a can of worms. I haven’t run a survey control network with GPS in a looong time. So, if you’re using the first Trimble R8 (2002) GPS models, um, best of luck. You’re going to need wide open access to the sky, no trees or buildings, and the times of observations will be critical to any GPS success.
Now, if we’re talking about R8-M2 or (2004) or R8-M3, R8-M3 or R8s, these are GPS+GNSS receivers. These bad girls pick up tons more channels and constellations. These GNSS equipped receivers changed the game. When I ran Trimble R6 and R8 GPS receivers, there were many times I had to pack them up. At certain parts of the day, generally in the afternoon, GPS satellite coverage dropped so low that RTK was impossible.
So consider that first. You also didn’t really clue us in on any project specifics. The ends ALWAYS justify the means and methods.
1. How big is the project?
2. How far apart are your control points?
3. Is there vegetation or buildings or any obstructions?
4. Accuracy needed - you asked if it was reasonable for 1-2 cm HZ & 3 cm V. That’s fairly sloppy for control, unless we’re talking GPS only again 😉
5. Post processing/baseline software? Plz tell me you have TBC
Maybe a RTS is the way to go. But, if the project is big enough to justify using GPS/GNSS, here are some tips I’ve learned from the school of Hard Nox.
1. Always use a fixed 2m rover pole
2. Measure base height multiple times - slant measurements can be rough and easy to mess up. Most folks don’t have a fixed 2 m tripod for their base, but that also helps immensely.
3. Observe base for a minimum of 1 hour. Do yourself a favor and crunch the static sessions. You can run them in TBC and/or OPUS
4. Run rapid static sessions with your rover. Multiple sessions, simultaneously running with your base, at a minimum of 15 minutes. Take several, crunch the sessions, and average.
You only mentioned 2 control points. Is that all you’re setting out? Again, since your GPS/GNSS experience is limited, I would search out some resources. There’s a ton of YouTube content these days. If you’re used to running control with a total station, turning sets of angles for control, it’s a whole different ball game. It’s quite easy to get an average from hard angles and distances. RTK and GPS/GNSS are finicky beasts - like everything is floating in space until you nail it down with multiple observations.
Cheers!
1 Good on you for doing your best to learn on the fly
2 Your employer (assuming its not you) should have a qualified person supervising this function and answering all your questions.
3 Off the top, if your intent is to check and adjust the control with a total station using a scale of 1, UTM is not the proper coordinate system to use. You should use a state plane zone at least or a low distortion zone at best.
PS I see you profile indicates Canada so scratch state plane. I'm not sure how much distortion is designed into UTM in Canada. The point is to use a system low enough to match the project expectation for surface measurement accuracy.
"Here’s one thing I’m curious about: what’s my expected relative accuracy when following this process? Should I just rely on the R8’s manufacturer specs (10mm + 1ppm for horizontal and 20mm + 1ppm for vertical), factoring in any leveling error from the rover? So, if I repeat observations on the same features over different days, would 1-2cm horizontal and 3cm vertical accuracy be a reasonable expectation?"
We recently had a thread about 1" vs 5" total stations and how the instrument spec rating is just one link in the error budget chain. Same goes for GNSS. The spec sheet ratings are what you can achieve under controlled laboratory conditions. That said, your 1-2cm target accuracy is achievable. I'd be occupying each point around 60-120 seconds, twice, and averaging.
This sounds like a big task for someone with only basic college course experience. The company you work for should have someone to lead this effort, unless they hired you expecting your college experience to have the necessary skills.
Honestly it sounds like you know of surveying concepts, but have no experience with successful projects. There’s a lot of specific details that go into what you’re wanting to do, and your result can be messy and wrong not understanding the data and software
1st set up the base to perform RTK and logging in your survey style. Set nail in ground set up base assign a point name from a here position. The receiver should be at to log the static data at an interval that will be consistent with OPUS if you are in CONUS so 30 seconds. I usually have mine set to 15 seconds. For any static baseline I might have to do. Base set up in clear blue sky. Run drive to another area that is within or around your site. Wide open set a point via RTK observed control point method maybe set one more at different location. These points do not need to be inter visible. After at minimum 4 hours maybe 3. Go back to your base end survey can be a different day. But move base to one of the other points. Set up rtk and logging set up on that rtk point same name. Rtk those other points. Same name store another. Now you have 2 observations that are all related and precision you can tell by the deltas abstinence datum who knows. If necessary the 3rd day or gap in time move base to third point and re rtk those other points. Or ones that do not meet the relative value you require. Again the 4 hrs gap means more at a tenth or so depends on equipment etc vs the couple hundredths in less than 3 hrs gap between observations. Your relative precision can be calculated easily now. Now all the Topo or mapping work todo in between all of the moves which can be a lot is all relative. To the first bases position. Once you send off the static files to OPUS example and move everything amgives you the absolute to the datum. Depending on equipment model. I bet your move is between a few feet and 15ft hz and vertical depending on where you are. Not 10cm. Although I have see. It close once in a blue moon. But usually not. Once you move to absolute check the other side static position for rotation and vertical to see if that is within reason based on OPUS results for a blunder out the gate. If say within a couple cm Xyz. Hold one do a least squares adjustment or hold both and apply the uncertainty to both the do least squares. Apply the correct weighting based on procedure and equipment etc. now you have brought the data all the Topo etc to a relative and absolute values considering any out liars etc to a point to know what you have. If all you care about is relative then that’s easy set the nail forget logging and relatively tie al points and follow the above minus the logging. Your relative is adjusted after outliers then you have your semi major and minor and true uncertainty along with a pretty well determined confidence in that dat based on weighting and such.
