I have been using my RTK iGage rover only connected to the INCORS network for years now. ?ÿI am happy.
The question is, whenever I decide to upgrade, should I switch to a base/rover setup?
Will I notice any (a) better vertical ability to stake and grade curbs (for example), or (b) better performance in tree canopy?
I can definitely say that I get better results, especially in the vertical, using a local base rtk setup versus RTN. We always use that method for highway mapping projects, with double occupations at different times of day.?ÿ
Relative position accuracy is dependent on how similar satellite propagation is to the rover versus to the base. Thus having the base near the rover has to generally help with iono/tropo propagation concerns. That would be especially true at times when conditions are changing.
Distance does little with respect to canopy and other multipath that affect only the rover.
Also a user of iGAGE rover?ÿ on NC VRS. Good results but wondering the same if I included a base in the system. Good question.
If I'm concerned about vertical I always use my own base and try to keep the distances down to 7 km or less, though I've stretched out to almost twice that far in a few instances on a 5 cm project.?ÿ I'd use CRTN (California) more often, but all the stations in my area are GPS-only, no GLN.
If I'm concerned about vertical I always use my own base
Have to agree.
We tried the local VRN.
But just kept on getting these 'odd' vertical inits.
Out by a random 100mm (== four inches) more than once too often
As "Any error rate high enough to measure is too high", we had to give up on the VRN 🙁
So now there's probably an extra half hour of my life spent on each and every city job establishing the base and traveling to benchmarks...
?ÿ
For GNSS I tell clients
+/- 0.01m HZ, +/- 0.015m V with own base
+/- 0.015m HZ, +/-0.02-0.025m V with network
Assuming good sky vis etc. And typically base/rover distances less then 2km in urban areas sure to radio range. Number's above approximate a 95 percent confidence interval.
- @jim-frame if your only doing control in good GNSS locations wouldn't make much difference being GPS only? Extra constellations only much benefit for more challenging points?
I often use RTK for recon and/or to get approximate datum. When you start stretching out RTK beyond about 10 km, it's harder to get the integers fixed with GPS-only. GPS + GLN fixes much faster and more reliably when going 15 or 20 km, at least with my equipment (Javad Triumph-LS + Triumph 2).
I disagree with your statement that distance does little when dealing with canopy. My testing has found that you can get very good results under canopy IF the base is close by, say under 1 km. Results degrade slightly the further away you go. But, I will admit that my testing of longer distances was before we had so many satellites and constellations, so maybe that has changed things. I have not tried longer distances since I got an R10-2 with ProPoint. I need to do some more testing...
I do use VRS a lot, just not on highway projects where vertical is critical. When I do use a local base, I try to keep it under 10 km if possible, and of course set the base out in the open even if it means setting a new point. Too many times I have seen base setups in less than optimal locations like this one...
+/- 0.015m HZ, +/-0.02-0.025m V with network
The term "network" is ambiguous -- it can mean single-base corrections over an IP connection, or a true VRS.?ÿ I don't have access to a VRS, so I have no feel for its accuracy, but I'm under the impression that its adherents claim greater accuracy over longer distances than with single-base RTK.
I'm under the impression that its adherents claim greater accuracy over longer distances than with single-base RTK
I would say this is generally correct, assuming a true multi-base solution and best practices with GNSS data collection. At least that has been my experience with the state network in Washington.
The only time it has gotten us into trouble is when we reactivate a legacy job and a crew goes back to a site that had been observed on a now-superseded datum the first time around. Outdated geoids can be the culprit too.
It is rare for sessions observed on the same network with the same reference coordinates to mismatch, but I have seen it happen before. However, the 2-3 cases I personally reviewed could be attributed to a bad fix rather than network accuracy.
Be careful - our state has a free network based on Leica and Leica software. We found that when using it with our Trimbles the vectors were not networked the same as when used with Leica equipment. They were coming off the nearest base in the network. This was several years ago so I'm sure that it's different now. But just be warned - your mileage may vary.
The way I see VRS working with Trimble equipment is that it creates a virtual base near the rover position, but actually stores it as a vector from the nearest physical base. This is done to enable tweaking of the positions, for example by computing a different epoch for the base and then using that, or a different realization.?ÿ?ÿ
Both are as bad as each other here. true nrtk on paper should be better but as Jim.cox points out earlier that 100mm "funnies" are not uncommon, something you don't see with base/rover.
I wonder if the base spacings in NZ should be much shorter due to our small country that has vastly different weather over short distances which the network solutions struggle to model.
The way I see VRS working with Trimble equipment is that it creates a virtual base near the rover position
What I've always wondered is how much more accuracy is obtained by using the virtual base over a single-base position to the nearest actual base.?ÿ I know there's some additional iono (and maybe tropo) modeling involved, and perhaps some orbit refinement, I just don't know how much that adds to positional accuracy at, say, 20 km from a real base.?ÿ (This sounds like a question for Gavin Schrock!)
A friend off line tells me;
Many of our users use the RTN to set up control, and break out a base if they need a little more vertical.
In response to your initial post, even using a base/rover combo your grades are still not going to be tight enough for curb layout.?ÿ Some people lay out horizontal with GPS and then hit their hubs with a level for grade, but no GPS manufacturer's equipment will get you the vertical accuracy you need for good curb layout.?ÿ Still need the robot or a level for that.
When I say highway work I mean aerial mapping for design.?ÿ
That sounds like how the Leica approach to NRTK is supposed to work vs Trimble's VRS which calculates a pseudo base position in the area your working and if you move too far it recalcs a new pseudo base position. Both use all the CORS in network to model atmospheric errors just have different methods of reporting "baseline" vectors