Relative Positional Accuracy

Good on you for trying to learn it. A lot of "surveyors" just throw some GPS at the job and if the distance look reasonable compared to the deed they just go with it, not giving any thought to RPA.

Continue to learn the underlying probability/statistics and related surveying error propagation theory. You should then pick a software to throw yourself at. I would suggest StarNet if you have no software otherwise, but there is a steep learning curve no matter what you pick/do.

There are other options: Trimble Business Center, Carlson SurvNET, etc...

PS: Some data collectors can give an estimated RPA if the point is observed twice. It has been my experience, though, that these values are usually over simplified and highly optimistic.

Just a couple of thoughts.

Some old school surveyors don't really understand nor this late in their career, want to learn a new bag of tricks. There are of course plenty of exceptions, but I've run into similar obstacles in my career. There is nothing preventing you from asking questions and learning. I feel it's your duty.

Two things about replacing corners lost to construction.

1. Were the corners properly tied before being wiped out or are you having to attempt to compute their most probable location prior to being wiped out. The first option is far and away the best. I've seen way too many corners put back where the surveyor thinks they should have gone and gotten it wrong. Sometimes by tenths of a foot, sometimes by substantially more. Having good ties ensures that the corner goes back where it was. A section corner controlling my personal property was reset by our local DOT 8' away from it's original location creating some real headaches for surveyors that follow.

2. Strength of solution. Having a single vector from your base using a single RTK shot to put a corner back is akin to a doing a conventional sideshot. There is no check and no way to prove the strength of your solution beyond your RMS values and PDOP at that snapshot in time. Any corners I set using RTK are followed up with static observations that include CORS data into the solution to verify the location using multiple vectors. The BLM requires multiple GPS vectors for their cadastral work.

Check out 'GPS for Land Surveyors' by Jan Van Sickle.

Good luck.

Willy out.

I agree 100% with everything that Willy said, I'll just add a couple of things:

- Trimble states a nominal positional accuracy for RTK in their literature, you can use that as a starting point. However, they also typically state the values at 1 sigma confidence, so you can also typically double them to get realistic numbers.

- If you have an R6 then presumably you have Trimble Access - the numbers displayed on the screen during an RTK survey are also stated at 1 sigma and so are highly optimistic.

- The only true test of a GPS observation is to take a completely independent redundant measurement. By completely independent I mean different setup, different constellation, different everything.

I have our crews shoot corners as Observed Control Points (3 min. obs.). I then have them re-visit the point after as much time has passed as is practical/possible, set back up with the pole oriented 180 degrees out from the first shot, and re-observe (as an OCP) using the same point number. The software will give you the option of averaging the two positions and will show you the standard deviations of the northing, easting, and elevation. Assuming both shots were "good" (within specification),those standard deviations are reasonable estimates of your probable error budget. Obviously the more times you can do this, varying the setup and under changing satellite conditions, the more reliable the solution will be.

One last thing - old school GPS guys were taught to "dump", or invert, the rod to cause loss of lock and to calculate a new solution. This is not an effective technique with today's receivers; if I had no choice but to shoot a point two or more times within a short span of time, I would end the survey and shut the rover completely off in between observations.

Good luck to you.... and yes, Van Sickle's book is a must for your reading list.

Previously located?

If you had located that monument with your RTK unit prior to its demise, you could replace it with little trouble.

Good on you for seeking knowledge. I am glad to see no flame activity..

As others have pointed out, a LOT depends on what you are working from. I try to avoid using RTK ties followed by RTK stake-out to replace corners with no independent check. Most of the time we reference the Corner with nearby accessories that will survive construction. When replacing the monument the local ties hold.

The biggest thing with RTK (IMO) is that ALL data collectors lie. Any expressed relationship between the point precision shown on the screen and your actual error budget is fiction. You are doing right to learn how to quantify and express 'real' results. The actual steps will vary widely depending on your specific network, equipment and processing software.

Reliable RTK stakeout can be done. Give us a bit more info on your network and I'll try to give you some practical advice..

Good Luck, Tom

Good for you Scatter to seek some input from what is likely one of the best sources of information on the planet regarding surveying issues. You will hear it all, and please thank our gracious hosts Wendell, Angel, and staff for the work they do!!

As a relative newbie to GPS (only since early 2000's), I was born & bred transit, then finally got an EDM, then eventually a total station with a data collector, now I'm into my best friend Mr Robot. Trimble GPS too.

One thing troubling to me in your subject line is "relative positional accuracy". It is where it is, and you either accept it or you don't. Sometimes people confuse "accurate" with "precision". Big difference. That's what we get paid for - to know the difference.

I know many surveyors who will not set a property corner with GPS. Construction staking for some things, ok. But never boundary lines. I think I've learned enough about GPS to know it's limitations and respect them. Listen to the others on this board and they can and will help.

I'm seriously thinking of selling all my GPS stuff and going back to the old fashioned static observations. In my area RTK can get real iffy due to the terrain. So what if we have 300 plus days of no clouds, those pesky mountains cause even more grief. So I'm thinking of dumping my entire tool box of RTK GPS and buying cheap static stuff. I trust Mr Robot because I can in a court of law say that "... yes I measured that line..". Can anybody say that after using GPS? No.

Best of luck to you grasshopper...;-)

As others have stated, the problem with GPS is that it always gives you an answer. Connecticut has some guidelines (a bit dated though). You can find them at ctsurveyor.com...on the right hand side click on codes and standards, scroll half way down and you'll see GNSS standards about half way down.

> I trust Mr Robot because I can in a court of law say that "... yes I measured that line.."

Well, yes and no. What your robot probably did is send a series of electromagnetic waves to a prism and determine the phase offset from a similar wave sent internally once it resolved the integer ambiguity of the reflected signal. There's a lot of clever physics and math going on between the time you press the button and the time the calculated (not measured) distance appears on the display.

That said, I'd rather have an EDM-derived value than a GPS-derived value simply because the error sources in the former tend to produce smaller uncertainties than those of the latter. But to call one a measurement and the other something else simply doesn't reflect (pun intended) reality.

Good one Jim. You kinda got me on the radial staking part, all though not explicitly implied. But I do generally physically walk all the perimeter lines of property I stake. Construction.... open world and just get it done, I suppose as I cross my fingers

My point is who in the world proves that GPS is any good for anything? NGS is who. Do we agree there? We accept their data on the premise that it is repeatable. We are banking on statistics that prove repeatability. Ok, so far it's worked for me... oh yea I can name one time when I was off a couple feet but my gizmo said I was spot on. duh

Then again, they keep changing geoids, moving the center of the earth, re-positioning space ships, and telling us "...you're now good to go...". So then we GO, locked cocked and ready to rock. Spooky stuff in my simple mind.

I just hope that our newbie poster doesn't think GPS is any better than that sliced bread in the cupboard with a spot of green goop growing on it.

Thanks for all of the quick responses.

First, a bit more background. The construction happened in 2007, by completely different people. During the resolutions of some of the easements, which seems to have drug on for 7 years, one requirement was the re-establishment of the corners and stakes representing the "proposed" easements.

We were assigned the job of the surveying and we did the property research, recon for existing monuments, the whole she-bang and based on our results, my boss directed me where to place the corners. This is entirely his call, as he is RLS.

Now I need to know how to determine a "RPA" for the new property corner monuments and actually set the corners in the field.

I have the R-6, a TSC-3 with Trimble Access, and we use a cell-phone tied to the internet giving us access to our state's CORS/VRS system for to use as our base station. I also have an R-5 base, but to be honest, I've never pulled it out of the basement. I do have a Sokkia Total Station which I could use to do a traverse, but I really would prefer to use the GPS if at all possible.

My plan was to stake out the corners, set the monuments, take a 3-minute reading on them and then return a few days later and repeat the process.

That would at least give me some information about the degree of error. However, I have no idea how to resolve that information (a vector for the error at each corner) into this RPA information I need.

I have Civil 3D, which can do least-squares adjustments, but it needs a true traverse to do so. I have no idea how to do that with GPS. My boss thinks, based on information from a continuing education class, that the data collector itself can give me the RPA. I know it can compute closure of a traverse, but again, I don't know how to do that with GPS.

I have seen, in real-time, what the data collector thinks is my horizontal and vertical precision, usually expressed in hundredths of a foot. I know it also records that information in the "RAW" file and I pull that later. Are these precisions the RPA?

Thanks again for all of the help and information.

I'll be looking for that book to add to my library.

Ok, last chime in on this thread.

I see some obvious issues.

1) You are out of your league. You don't know what you're doing and you've BS'd your boss into thinking that you do. That is a problem of yours we can't fix. Get some training.

2) The best ways to learn is to make mistakes. Then you learn not to repeat that. Oh gosh, there is that GPS "repeatability" thing again. Go ahead and goof something up. I have, and I think about 99% of the experienced surveyors on this board have too. Lots of jokes about our goof ups.

3) Think about the end product before you ever set foot on the project. If you see that, then you do have hope. If not, go work at Home Depot. I assume you see that end though, so think it through. Just pretend that you are signing the final plat that may end up in court.

4) Last one and I may get beat up here by other folks. Least squares, compass rule, or any other form of "adjustment" does not fix one single thing. You get warm and fuzzies by fudging numbers, but the point is where it is. I've never adjusted a traverse, level loop, GPS gizmo stuff, or even pin cushioned a corner in my life. Every property corner in the ground with my number (MI & AZ) (thousands of them) I personally set.

It is what it is. If you don't like it, then go back and re-think it again. I've found that 90% of the time I thunk something was amiss, it was. I found more evidence to make my original findings perhaps not so right. Then I made the necessary adjustments (in the field) and got my warm and fuzzies and got to give momma a big hug instead of fretting all night. Kinda not really billable hours, but I do sleep better and momma is happy.

@Wayne:

You've made some assumptions about me. Wrong assumptions. I never convinced my boss of anything to get my job or do the things I do. I've been an Engineering Tech for nearly 25 years. 3 years ago, by boss gave me the equipment and told me to use it. End of that. I've been around Civil Engineering, Surveying, Geometry, property plats, subdivisions, infrastructure, drainage, the whole bit for all of that time. That doesn't make me an expert. I've never - not once - told anyone that I was an expert or a surveyor.

I knew that I was out of my league, that's why I came here asking for help - thinking I would get a little assistance instead of the "home depot" speech.

I plan on learning whether I get the information here or not - I WILL find my information and get my experience and do my job.

I'm not looking to "tweak" my work. I'm trying to find out how to provide the information that I need to and produce proper work. I currently don't know how to do that and I have no one to ask. This is why I came here - thinking that proper surveyors could explain it to me. I'm not asking for a cheat, I'm asking for information I can't find anywhere else.

The standards in my state say that plats done with GPS must state their "Relative Positional Accuracy", I don't know how to find that to report it. I don't know how to generate it.

I've made mistakes. I've learned from every one one of them. I will make more - you're making me think that coming here to ask those more knowledgeable than me is one of them.

I'll leave you with one last question: If those that want to know can't find out, how can they ever be expected to do something?

Some people can't get through a day without trying to step on somebody...

anyhow.. Follow the advice to check out the standards link. After that try to glean tidbits from those willing to help. Test everything. Some of us are full of crap (all of us at some time or another).

My practical advice would be dust off that base. With a little practice it will be up and down in minutes. Your solutions will tighten way up. If you do second occupations try getting them the same day at least 2 hours later. That will get you a significant constellation change. Doing it the same time the next day is like back to back shots (nearly worthless as a check). Check in every time you initialize. It will save your butt.
heading to bed. Enjoy your quest and good luck, Tom...

I check everything with double redundancy.

By that, I mean that I check my position from two different control points and compare the results.

Research and find an existing plat that shows what someone else has done to meet the requirements.

😉

In order to compute RPA or RPP, you need redundant observations, as the others have said.

I have only done this with a base-rover setup, but IF you can extract vector data from either your virtual reference station or CORS to your measured points, then these are my suggestions:

If you want to compute it per the latest ALTA standards, you need to have software which will compute relative error ellipses between all the points which control the boundary of the property surveyed (property corners, as I read it). Star*Net is your best bet for that.

BUT based on your reference to vector closures and RPA, it seems your local standards may be more in line with the older ALTA standards. Florida DOT has free software called VectorNT ("Hector the Vector) which will compute the vector closures and positional tolerance based on a minimally constrained adjustment. It even formats your data for you. Using this would in my opinion be the easiest way to have a defensible argument that you have met the standards. Your procedure would be to measure each point at least 2 different times (at different times of day - hopefully this gives you vectors from different VRS stations), extract the raw vector data from the raw file ( make sure it is in meters), enter the data into VectorNT, following the program documentation, adjust (minimally constrained), run the GPS loop-closure and positional tolerance routines, and if all is well, you are done. Start with maybe twice your displayed precision as your error estimates in the input file, and make sure the adjustment passes the chi-square test. You can do this at home on your own time if you have to - when you show your boss that it can be done without any great additional expenditure, he will be more likely to make it part of your standard procedure. A word of warning, however - achieving good RPA with RTK can be iffy. You may not pass, and then you will need to add some conventional measurements, and use paid software to get the results you are looking for. Also, don't necessarily use the results of your minimally-constrained adjustment to determine the final coordinates of your points.

The most authoritated answer to your question is the "Federal Geographic Data Committee Part 3: National Standard for Spatial Data Accuracy". The Duncan Parnell web site has a short online course and a spreadsheet for computations using the above stated standards. I believe it was less than $50.

:good:

>
> 1) You are out of your league. You don't know what you're doing and you've BS'd your boss into thinking that you do. That is a problem of yours we can't fix. Get some training.
>
> 4) Last one and I may get beat up here by other folks. Least squares, compass rule, or any other form of "adjustment" does not fix one single thing. You get warm and fuzzies by fudging numbers, but the point is where it is. :-S I've never adjusted a traverse, level loop, GPS gizmo stuff, or even pin cushioned a corner in my life.

I am not looking to beat up on anyone but methinks you should take your own advice and get some training on the issues italicized above.
Possessing a survey licensee does not come with some survey knowledge providence bestowed on the licensee. Like everything else in life, it is a continuous process of experience, learning and adapting to the current time. It is not static.

> Possessing a survey licensee does not come with some survey knowledge providence bestowed on the licensee. Like everything else in life, it is a continuous process of experience, learning and adapting to the current time. It is not static.

Thanks Robert for some very sound comments, which I take in respect. However I do suspect that our certain Mr Brain is looking for input for which he is not qualified to evaluate. My opinion and nothing more. When somebody asks for advice, they should be prepared for answers they do not like.

Me, I've only been doing this crappola for going on 40 yrs. I liked the world better when it was flat and there was no GPS. But it's not, so I'm a dinosaur in the making.

PS - I assume that the word "not" should have been excluded from the above sentence you wrote.