One reason on not using cell is cell coverage and the other is latency. I don’t mind using cell and have done it both with single base and VRS FKP etc. In rural areas cell coverage gets spotty. In good cell areas certain times of the day usually when kids are getting out of school and the tower can’t handle all the traffic and we get latency which intern can give us suspicious data or less than desirable positions.
This has been our experience by and large. That and we would have to get an external modem because the receiver internal modems are all 3.5G. We often see high latency even in large metro areas. Our RTN is full-constellation, and even running RTCM3.2-MSM it's about a 50/50 chance whether we see all four constellations in the solution.
We also work in enough out-of-town areas (at least my office does) that it is almost always easier to just run UHF corrections from the start.
We're a larger firm, and corporate is loath to pay for anything that isn't being used day in, day out.
We use a microhard bullet. I originally bought it because it has a serial port and I was using that with an R7 as a base (which only has serial ports). Now I use it with an Alloy as base, or sometimes R10's. If you get one with wifi then the R10/R12 units can function as a wifi client, whereas the Alloy connects directly to the ethernet port on the modem. If you have a static IP then it is easy to connect directly to the base receiver. If you don't get a static IP then you can run SNIP on a computer in the office and send the corrections there, and then connect to SNIP from the rover.
I recently found a great option...simbase.com. You can get a sim from them that will work pretty much anywhere in the world. They charge $0.01/day for service, which you can enable and disable anytime from a dashboard. Data is charged at $0.01/MB, which comes out to $10/GB if you use it a lot. A public fixed IP is $0.30/day, which is $110/year. I was paying AT&T $56/month for a static IP sim with 5 GB, but I never used anywhere near 5GB, and there were months I didn't use it at all. But I like always having that option active. I switched over to simbase after testing it, and it works fine.
As for latency, I have not noticed it at all, but we almost always are doing static occupations (2 to 3 minutes), not anything kinematic (i.e. moving while collecting like in a vehicle or walking and doing continuous points).
My comments are based on these assumptions and 45 years of experience in route survey:
1) There is no need to replace the centerline vertical component. If this were a requirement the comments would be much different.
2) The horizontal centerline component is being identified by a visual estimation. (eye balling) This means the location may be off a few tenths of a foot or more from where it may actually be based on a centerline survey using centerline control monuments from as built plans. This of course is the proper way to do this which there is no time for as stated. I would have a problem with that if I were approving the work but luckily for you I'm not.
Before RTK the easiest way to preserve estimated centerline location was to simply tape offsets to hubs set out of harms way and pull it back as needed. This would still work in a pinch. This may still be a consideration all things considered. Particularly for an inexperienced RTK user.
To use RTK start in the middle and work out both ways setting control as far from the first base as the radio link will allow. Set your control on hilltops so as to cover the valleys. The first base position coordinate can be assigned or approximate. After additional base points are set and occupied check back to the base point it was set from for your warm fuzzy. Because your centerline locations are estimated to start with your concern for RTK accuracy is a mute point because your estimated locations are well within the tolerances for fixed RTK survey.
With interstate pavement, the centerline seam is vitally important. Usually after the mill there will be a straight side, normally the parking side and the seam is laid out from it simply pulling a tape over. There will probably be two lifts. Those lifts should overlap so the first layout will be maybe a foot past the seam and. The final seam should be packed very tight and the goal is to make the painted centerline on top of it. The idea is to keep as much traffic from running along it since it's the weakest part of the pavement. I'm assuming you were asked to locate the existing seam and then lay it out. One caveat is that design comes into it when there are things like turning lanes. In those situations you may need plan sets for more complicated pavement layout.
the interesting thing is there are two jobs on this interstate separated by 3 miles. It is a 3 lanes, southbound interstate. The engineer on the first job wants to collect the yellow line next to the median as the controlling feature. On the second job, the engineer wants to collect the crown which is between lane 1 and 2, Lane 1 is next to the median. I will probably collect both features just in case. The goal is to collect the information and than will use them to lay out the paving widths with a Tiny MobileRobot. Its great to hear all the answers. They all have information to use.
Yeah I say that for establishment of starting control the VRS and cell is awesome when it has been tested. But once you are on a site and doing mapping boundary stakeout and more control for relative precision on site you can’t beat uhf. I have had a few jobs that I lost radio on and I already had my base logging data no cell area at all. So I just did fast static while looking for prop corners and doing some sketching. Processed the baseline that night set base on it the next morning and logged another redundant fast static at the old base location. And rtk some additional points to tie that side of the project together. Trimble makes it easy for sure using multiple tools on a site. Cell is great when it works but it can bite you for sure.
That’s not to bad. On 3 miles. In between each project. What’s the length of each project though.
5-10 miles
