Activity Feed › Discussion Forums › GNSS & Geodesy › VRS vs RTK Base Station (Radio)
-
VRS vs RTK Base Station (Radio)
Posted by JimMac on March 4, 2022 at 3:30 pmHi Group, ive been looking for a Survey Forum for awhile and I’m excited to be here! I assume this topic has been discussed in length, and i will search for previous discussions, but wanted to get me question out there. ive used VRS (Can-Net) back in 2010 to 2012 and now again for the last couple of years. What i found back then and still now is that we have better vertical consistency with the Conventional RTK Base setup. Typically completing layout within 1km of the setup. But pretty much every report i find online comparing the 2 says that i should have better results from the VRS. In 2011 i observed really bad vertical from Can-net at various times of day (but with good cell coverage) and locations (spotty cell coverage). It was explained to me then by a Cannet Tech, that i needed to keep an eye on the signal latency, and if it started to get much over 1 second then i was going to see degradation in my vertical. I accepted that as makes sense, if my corrections are delayed then my position should be effected. In later discussions with my SP’s techs when i was considering subscribing again, i brought up this latency concern and they didn’t know anything about it, and seem to think that this wasn’t a ‘thing’. Currently our crews are observing a 20mm tie in in the morning and a >35mm tie in the afternoon with the VRS, while consistently tying 20mm to 25mm with the conventional Base setup. So my question is can anyone confirm the signal latency effect on vertical or has seen the same VRS vs Base station variance.
Thanks
OleManRiver replied 2 years, 6 months ago 10 Members · 14 Replies -
14 Replies
-
With my limited understanding of the differences between VRS vs. traditional base/rover RTK, I’d like to know what the satellite positioning was like during the times of your largest errors. The mentor that taught me most of what I know today about GPS, once stated that for the tightest vertical, you needed at least one satellite on the horizon, preferably two. He equated it to double chaining, if the control from which you are chaining is close together (being straight overhead), your errors will be greater than if your control was 90 degrees from the target.
My experience with checking into established control with VRS has been great in the midsouth. Our vertical error between the published coordinates and the collected data is usually under 0.04′. On a rare occasion, we get errors up to 0.11′.
Just a thought.
-
@kevin-hines Hi Kevin, thanks for the reply. I haven’t looked into constellation geometry, i considered it a constant as we have had both VRS and Base station operating at the same time on site to observed the variance. We currently have our Site inspectors equipped with the VRS so they can do quick asbuilts or replace the odd stake, and our crews run base/ rover for layout.
-
Yes, vertical errors are greater when using network VRS versus on site base RTK. Signal latency from cell corrections is greater than on site radio, thus positional uncertainty is greater. We have this discussion all the time with our grading crews in the western part of the state. They are used to having on site base corrections and we are transitioning to VRS corrections on all bulldozers. The vertical error is greater, but we have less equipment to worry about and also can remotely add files and solve issues on the dozers when they arise.
-
Posted by: @jimmac
Hi Group, ive been looking for a Survey Forum for awhile and I’m excited to be here! I assume this topic has been discussed in length, and i will search for previous discussions, but wanted to get me question out there. ive used VRS (Can-Net) back in 2010 to 2012 and now again for the last couple of years. What i found back then and still now is that we have better vertical consistency with the Conventional RTK Base setup. Typically completing layout within 1km of the setup. But pretty much every report i find online comparing the 2 says that i should have better results from the VRS. In 2011 i observed really bad vertical from Can-net at various times of day (but with good cell coverage) and locations (spotty cell coverage). It was explained to me then by a Cannet Tech, that i needed to keep an eye on the signal latency, and if it started to get much over 1 second then i was going to see degradation in my vertical. I accepted that as makes sense, if my corrections are delayed then my position should be effected. In later discussions with my SP’s techs when i was considering subscribing again, i brought up this latency concern and they didn’t know anything about it, and seem to think that this wasn’t a ‘thing’. Currently our crews are observing a 20mm tie in in the morning and a >35mm tie in the afternoon with the VRS, while consistently tying 20mm to 25mm with the conventional Base setup. So my question is can anyone confirm the signal latency effect on vertical or has seen the same VRS vs Base station variance.
Thanks
I don’t know the answer to your specific question, but I can confirm that our crews have found base/rover setup to be typically be more accurate horizontal and vertical. Usually we’re just using it it for horizontal anyway.
We’re also on cannet.
-
I think for dozer work VRS would be plenty accurate enough.
-
In my opinion the accuracy is definitely better with a local base RTK than with VRS. We usually do the local base RTK over cell, so there may be slightly more latency than if we were using radio, but not really noticeable.
Another issue is whether the VRS is using more than just GPS and Glonass. Our base and rover are both GPS/Glonass/Galileo/Beidou. Some VRS we use are all 4, others are just two constellation .
-
We have really good results with Base/Rover, Geoid 18. I’ve never used a VRS but 20-25mm would be unusually sloppy for my Base/Rover set-up.
-
Posted by: @va-ls-2867
Yes, vertical errors are greater when using network VRS versus on site base RTK. Signal latency from cell corrections is greater than on site radio, thus positional uncertainty is greater. We have this discussion all the time with our grading crews in the western part of the state. They are used to having on site base corrections and we are transitioning to VRS corrections on all bulldozers. The vertical error is greater, but we have less equipment to worry about and also can remotely add files and solve issues on the dozers when they arise.
This is correct. We see the same thing here in WA. In some areas where we have excellent cell service and latency is low, positions are tighter. Others not so much. Vertical repeatability is definitely better with base-rover.
The other part to remember is that some RTNs are not full-constellation, whereas recent base-rover setups usually are, so more SVs will improve the solution as well.
Also, if you are running with a single-base mountpoint (not a true network solution) the position will be subject to the same distance-dependent relative error as a base-rover setup, so unless you are within a comparable (base-rover) distance of that base station, you’re going to be seeing more error.
“…people will come to love their oppression, to adore the technologies that undo their capacities to think.” -Neil Postman -
Posted by: @va-ls-2867
Yes, vertical errors are greater when using network VRS versus on site base RTK. Signal latency from cell corrections is greater than on site radio, thus positional uncertainty is greater. We have this discussion all the time with our grading crews in the western part of the state. They are used to having on site base corrections and we are transitioning to VRS corrections on all bulldozers. The vertical error is greater, but we have less equipment to worry about and also can remotely add files and solve issues on the dozers when they arise.
Wow, that’s interesting. Never seen a fleet of machine guidance systems using VRS/RTN exclusively. We have found that vertical error on VRS/RTN is out of tolerance when Network Base is farther than 5 miles from the project site. Also, you are 100% reliable on a 3rd party (usually a government agency) to provide GNSS corrections, and also dependent upon cell service.
-
Latency is the most probable cause of the variation you are seeing.
The quality of an RTN depends heavily on their diligence. Cellular connections are so simple to install and get online that it is tempting to use them entirely. The time for a signal to get to the RTN server is typically several times longer than one using a LAN. An RTN operator should try to get his stations connected to a LAN and resort to Cell service only when necessary.
The “tightness” of results will frequently be better with a local base station but is the result of proximity and no latency. The “accuracy” of results also relies upon the base station that may or may not have been surveyed in well, as well as a local base is subject to any interference that exists locally. A VRS from a reputable RTN can be nearly perfect, in this regard, being created from several CORS that model the atmospherics and it lacks local multipath and blockage.
Distance from the base is also an issue. Typically, the precision degrades at about 1ppm of the distance from the base. The VRS is usually set up to follow the rover around keeping the baseline length nearly zero.
Bottom line: the care and quality of the system used dictates the quality of the results.
JAC
-
@john-hamilton “We usually do the local base RTK over cell,”
Thats really interesting, im currently looking at setting up our own permanent bases around the city and was looking at the Trimble R750, which doesn’t have an internal radio but does have a cell module. I assume you require a carrier and is the correction sent ‘directly’ meaning no internet involved?
-
We have 4 R10’s, and kept sim cards in each one (only one with a static IP to use as a base). Since AT&T shut down 3G two weeks ago I had to redo my setups since the R10’s do not have ethernet ports. I now use a microhard modem which has wifi and a static IP. I set the base R10 to connect to the modem as a client. Then I had to configure the firewall in the modem to pass the port numbers through to the R10. I setup one port for CMRx, one for RTCM3.1, and a third port number for web access to the R10 interface. The latter is nice because I can then connect to the R10 interface from chrome on my phone.
I also sometimes use an alloy as a base, it does have an ethernet port so I can use that directly connected to the modem (a different microhard modem that does not have wifi). I originally bought the microhard modem to use with an R7 transmitting corrections over serial, both modems have serial ports.
Yes, you do need a static IP, OR if not you can send the data to an NTrip caster, I use SNIP running on my office computer. We have had long linear projects that require several base stations, SNIP will automatically send corrections from the nearest one. In that case we had a VRS base at the north end, I connected SNIP as a relay to that as a single base, then set our units 1/2 way down and one at the other end. So SNIP would pick the closest of the three to send corrections.
SNIP also has a free NTrip caster…rtk2go.com. I have not tried that but I assume it should work just fine for a single base, and you won’t need a static IP. You can get AT&T sims with dynamic IP for $20/month from dataactivationservices.com, and also a verizon sim with static IP for the same price. I have a sim with static IP from AT&T, it costs about $55/month. When I had the AT&T static IP sim in my R10 it worked fine as a base in Guyana last year ($10/day), but there were a couple of times the data cutout for a few minutes, I believe they were blocking my constant stream when their network got busy.
-
I would say that base and rover has better relative precision than a vrs. Many sound advice on why this is. Latency vrs network spacing etc. However with good procedures vrs can be more accurate. Even though the precision is higher . Two different things that collide together. One of the best advice I could give anyone doing rtk and concerned about vertical is to teach the crew to look at the numbers of the satellites themselves and when vertical accuracy is needed make for sure you have a gps satellite above you and at an elevation of 65 degrees or more period. Watch vdop and use a gap of time so constellation changes i like 4 hours and I don??t sweat more than a tenth of a foot between the shots. The average is closer to the truth than many shots repeating better in a closer time frame. Note GPS satellite not glonass. As a geodetic orbital scientist watching monitoring gps daily 24/7 365 per year and combining my survey knowledge its a no brainer. Now if i am concerned with relative precision my approach is different. Accuracy vs precision. NGS has a rtk guidelines and use to have a document showing rtk shots every hour and if you study that it will lead you to where you should go depending on what you are trying to achieve. It is insane what can be done now with rtk equipment. I mean you can get results like we were getting with static many years ago if procedures and understanding are combined. Now static today can achieve tremendous results it is just not practiced as much now days. Time is money. Lol.
Log in to reply.