GPS for Traverse Control

Well, since you asked, here's how we do it. We will set a pair of traverse points at each end of the line that needs to be traversed. If we have the time, they will be static control; however, if we don't then they will be RTK observed at least 4 times over the course of the day (maybe the guy who process' the data isn't around to process the static control or a lot of other reasons). Static control is the best.

Anyway, after they are set, those values are loaded into the dc and the traverse is run on the surface keeping up with HI and HR. So we have grid points and a surface traverse that we close out into two other grid points.

THEN when we reduce the data in the office, I set the parameters of my zone in my job in Carlson and let it calculate the CSF at each point and apply that to the surface ties. Then Carlson lets me see how well we hit the grid points since we are now comparing apples to apples.

My thoughts are if you start on the grid, keep it on the grid.

If you do not have a lot of relief I would create an LDP for the site with the origin at the center of the project.Then you just convert your grid coordinates to local and go.

Thanks for the response,
So Carlson adjusts the ground coordinates to grid? Or does it just let you see how the check was and leave the coordinates ground? If the cords change to grid what do you do when you go back and layout from the grid cords?
Thanks again,

How do you adjust grid to ground? I thought scale factors were only for distances not cords?
Thanks for the help.

I like to scale all the GPS from one point. Use the scale factor at that point as determined by your processing software, and verify the value by checking Corpscon and any local NGS control to make sure there aren't any ambiguities. Then scale up from grid to ground using 1/CSF.

I wouldn't worry about the bearings, because your just expanding everything out uniformly, so they will remain the same. Then you'll be on ground, which your total station is measuring on. Using this method, you can always go back to that scaling point that you originated from to scale back down to grid if necessary.

Least Squares Adjustment. Properly weigh the traverse observations. Then run an adjustment. Examine the residuals. If they look OK, then use the coordinates from the adjustment. I would much rather know that all the points in the traverse are accurate to 0.015 m for example than to know that they are part of a 1:10,000 traverse.

Sorry I am new to this... To Scale from grid to ground I divide the grid N by CSF and grid E by CSF?

Grid to Ground

Grid Northing/CSF
Grid Easting/CSF
Equals
Ground Northing and Easting?

> Thanks for the response,
> So Carlson adjusts the ground coordinates to grid? Or does it just let you see how the check was and leave the coordinates ground? If the cords change to grid what do you do when you go back and layout from the grid cords?
> Thanks again,

It applies the appropriate CSF to each traverse station and the surface distance measured, making your surface measurements, grid measurements, and subsequently, grid coordinates.

The method I'm describing would be CAD based. I.E. grabbing the GPS points only, choosing one as the base point and expanding (scaling) them out by 1/CSF. So if your scale factor is 0.9999 you scale by 1/0.9999 or 1.0001 which adds about 1/10th in 1,000ft. That's a rough approximation of an average scale factor in NJ.

So your earlier thought was correct in that you are scaling the distances between the points, which will result in the ends of those lines having new coordinates (Ground Coordinates). Then you should be able to see how close your traverse distances tie in with your GPS. Without scaling, the lines between your GPS should be too short. If you want to scale just by coordinates, there are ways to do that as well, but it's more complicated.

Let me describe the method of several of my more complicated projects. One was a 5 acre survey, where I kept running into neighbors that wanted or needed surveys. When it was all said and done I did 7 surveys and tied into 4 minor sudivisions and 2 major subdivisions. I was well into the first 5 surveys when it became apparent I had to reach out for adjoining information. Essentially 2 20+/- acre lots that became subject to subdivision. I surveyed both and turned data over to others to divy up. Very wooded area and 99 traverse points when all was said and done.

I started with 2 GPS points on existing traverse points in a field on a 10 acre lot to the East of my PQ and 2 pair at either end of a proposed traverse along a road to the North of a subdivision North of the first add on 20 acre survey and a pair to the West on another road. From the most NE GPS pair catching a few sub monuments West along a road to the other GPS pair thence past the NW corner of that 20 acres and through the woods eventually tieing into my first 5 acre lot. Then starting at the GPS pair to the West through the SW corner of that 29 acres and through swamp to a corner to the 7 acreas adjacent to the first survey. Then I adjusted it all.

When the second 20 acre tract was surveyed next to the first 20 acres I started at the previous West GPS pair southerly along that road then East along my original frontage road to the initial area. I then GPSed 3 of the most visible traverse points on that route and 2 pairs on monuments, one pair each on adjacent major subdivisions to the West. I did a limited adjustment on that road traverse. From the 2 pairs in the subdivision I shot some additional sub mons on separate mini traverses. Those were used to confirm some sub mons along that PQ. I was satisfied with using remote GPS points w/o physically ttraversing them in.

On future large farm surveys I used triple GPS points to tie in remote property corners and roads to the main survey. By traversing through 3 you get a better check on your field work.

Generally the time budget does not increase very much for 3 points versus 2, and the confidence number is much, much higher. With 2 GPS points compare to a traverse you have distance only. With 3 GPS you have distance, angle and distance as well as the geometric distance check. In a few cases I have only occupied 1 of the GPS points with the total station, sideshotting the other 2 GPS points and from 1 to 3 monuments. My main GPS point in a set is almost always an OPUS or long OPUS-RS while the others may be shorter OPUS-RS or L1 only. With my mix of equipment I can personally cover up to 3 L1/L2 and 6 L1 simultaneously. Usually it is less work to haul less equipment and split occupations.

I bring my GPS coordinates into Carlson and usually scale them about my best OPUS position. It is then possible to bring in later GPS points and scale them about the same control. Remember not to scale the elevations.

Paul in PA

Just to add to what I said earlier, adding a number to the coordinates provides a shift, and multiplying scales it. If you multiplied the coordinates by the scale factor, I believe your distances (point to point inverses) would then be on ground, but you would have moved all the points by that factor. Then you would no longer have a point that was common to grid and ground systems.

To do a real coordinate based scaling procedure using excel or some other spreadsheet program is far more complex than just multiplying the values. I just do it in CAD, so I can visually see the change on the screen and then compare it to the total station points. Hope this helps 🙂

Here is a recent example.

Three of the perimeter traverse points and one interior point near the middle had GPS on them. Forced centering was used. I would have driven myself (more) crazy by trying to adjust these traverses by any means other than least squares. Here are the station confidence regions based on fixing two CORS and a nearby benchmark.

the beauty of an LDP is that there is no need to apply scale factors.

what softwares do you use? field and office?

I do the same as Kris, keep it all grid or all ground, mixing and matching is a recipe for something to get screwed up.

SPC, UTM, LDP, all the same procedure, some will have smaller convergence angles and less scaling, procedure is the same for all however. It is possible that grid and ground are essentially the same in all of those, more likely however in a properly designed LDP.

Be prepared to find that the conventional traverse won't fit as well as you might hope for even when ALL is done correctly, especially on long linear traverses as this is the situation usually being controlled by the procedures you want to use.

SHG

Grid to ground or ground to grid is a cogo thing, rather than a cad thing.

Yes "Scale Points" Is A Cogo Tool

I scale my GPS points to ground. Then set my field surveyed points over the scaled GPS control points.

Paul in PA

Run traverse, making sure that enough marks are in the open to hit with the GPS
Survey as many points as feasible with gps, ensuring good coverage along the whole length of the traverse.
Scale gps to ground using the average traverse height and easting (assuming your scale factor corrections are similar to ours, based on utm)
Helmert adjustment of traverse data to GPS.

If significant residuals, then look at reprocessing the total station data using the gps as control, but in my experience so far that hasn't been required.

I do this on pretty much all of my projects. The first thing to remember is to keep everything on the grid until your network is adjusted. If you try to scale your GNSS control to the ground prior to the adjustment then you will add that distortion to any error in your conventional control. Most modern programs will adjust the ground distances to grid on the fly. By adjusting on the grid the resulting errors will be from the actual measurements.

That being said, my standard procedure is to establish a base line utilizing GNSS from which to start the project. As has been said by others on this link, static observations are best for the final control. With RTK you are potentially looking at 10mm + 1 ppm error in horizontal as opposed to 5mm + .05 ppm for static. For my purpose I usually collect RTK data along with the static data on the first baseline to seed my conventional network. Then depending on the length and layout of the control I will establish additional static GNSS control points approximately every 1.5 or 2 km and at least one at the end. I adjust everything utilizing least squares with cross ties and have found that an azimuth pairs are not actually necessary for all networks as long as you have enough fixed GNSS in the mix (this is probably good for another thread). The nature of my work creates a lot of long sometimes 10 or 20 km in length and this system works great for them. The main thing to remember is scale to the ground after the adjustment is done.

Hope this helps.

John