Off the cuff here??I??m a skeptic??GPS is NOT for woods??but check this out??Couldn't help but share this:
I have been working with a Javad Triumph-LS and Triumph-2 base for about 9 months now. When I purchased the units I went around some forums looking for some idea of the repeatability that I might expect from this thing while doing wooded boundary surveys. I found a lot of happy users, but not really any data to compare. I wanted to post a real-life scenario that I just completed so that others can have an idea of the results.
I am a private surveyor??not a sales pitch here??just a scenario with facts. I have been a diehard Trimble user for a LONG time but as a small firm surveyor working from home, I just couldn??t swallow the R10 price. Hence the LS purchase??one year??already paid for??
The following is data from a 20 ac. wooded boundary I just finished?? Actually, I started back in April but it was for a friend and I figured that it would be a great project to try Javad out since timing was no issue. This was some heirs property that had never been resurveyed since it was conveyed in 1910. None of the deeds agreed all that great??but the corners were somewhat well marked. Back in April, we took the Javad-LS out there and in one day located well over half of the property corners??the problem being??the locations didn??t really fit the deeds too well??not too surprised??but, was it the LS?? Or was it the deeds?? I really had no indication on how good this thing was? Well...long story short, I let the survey sit on my desk until last week??had a short lull here between the holidays so we went back out there and in three days cut / traversed the whole thing conventionally. Then I went back and located a few more corners with the LS so I would have a comparison all the way around this tract. I??ll describe a quick field procedure here??then the results:
For the Javad, it was a base / rover RTK survey, radio link via mobile hotspots. Base in a nearby field??90% clear sky about a mile away. Each rover boundary location was measured with Javad??s ??boundary? survey profile which goes through at least a 3min verification process??please understand that some locations took up to 20 min. and some locations would not fix and had to post process for a fix.
For conventional, I used a 3? Trimble S7 robotic total station. Each traverse turn I measured two sets of rounds from poles/bipods. Side shots I measured one round letting the robot lock on the glass the whole time??typical medium accuracy (normal) woods traverse. The closure turned out fine: 14 turns - 5? angular & 1/33514.
For adjustments, Javad LS does this thing called ??DPOS?. It processed the base to rover positions using the static data collected while on each boundary point. It then back in the office it automatically processes the base using CORS data and applies a shift to the base just like holding an OPUS position. (two separate base sessions 8 months apart agreed within 0.006????guess my pole is plumb!). You can view all the post processed coordinates next to the RTK derived coordinates and review the respective RMS/epochs/stats??and choose what you feel is the best coordinate for any given location. Last step in the LS was to scale the project to ground and output the data...both generally easy steps in the LS.
For conventional, I typically run a least squares adjustment through Trimble Business Center??but so far the Javad does not play well with TBC??or let??s just say...that is on my list of ??to-do??s?. In this case I went with a compass rule adjustment from Trimble Access??clean as a whistle??and it applies the adjustment to all the side shots??
Now I have one complete adjusted conventional boundary data set, and 19 common corners with the Javad Data set.
Combining the Data: In Carlson survey there is a coordinate transformation routine that lets you best fit two local coordinate data sets together in one least squares smack??great routine...it outputs residuals and lets you choose what you want to turn off or on in the equation. Below are the results of the least squares comparison of the conventional traverse to the RTK measurements. In all fairness...These are ALL the common points??all ??yes? for processing in the least squares comparison??no outliers that I??m leaving off here to make this thing look better than it is??this is the raw comparison of the two data sets as measured and described above??check it out for yourself.
Results:
I thought this would be important for anyone to see who has the same questions I have...again...(Disclaimer!)...this is not a miracle machine...but it did WAY better than I expected...so I was compelled to share.
Also attached is an aerial showing the project location and the corners that posed a difficult RTK scenario. Lat/Long if you want to zoom in on the site on your own.
~ Timmy G.
I'm going to need to spend a little time with your post. More than I can at the office.....but you have spent a lot of time on it and it is appreciated just for that. A terrific way to introduce yourself.
I did not know that Carlson had a transformation routine like that.?ÿ I need one of those in my tool kit.. I have devised a way to trick StarNet into doing something like it....but that looks much more slick.?ÿ?ÿ
What experience do you have with other modern GNSS setups? No doubt Javad works better than older GPS only rigs, but how does it stack up against modern competition in your experience?
so far the Javad does not play well with TBC??or let??s just say...that is on my list of ??to-do??s?
Here's the procedure I use:
1.?ÿ Make sure you have JPS2RIN installed on your computer.?ÿ It's available without charge on the Javad website.
2.?ÿ Select the .jps files (the Javad raw data files associated with each point) in Windows Explorer.
3.?ÿ Right-click and choose "Convert to RINEX 2.11"?ÿ This will create RINEX files for each of the selected files.?ÿ The RINEX files will be created in the same folder as the .jps files.
4.?ÿ Import the RINEX observation files into TBC.?ÿ These are the *.18o files (for this year; next year's will be *.19o).?ÿ You'll also need to import the *.18N (ephemeris) files, unless you plan to use precise orbit files.?ÿ The import can be accomplished via drag-and-drop or through the TBC Import dialog.?ÿ (If you use precise orbit files -- which I usually do -- you can get the .sp3 files from the IGS website.
5.?ÿ The import will bring up an observation check-in dialog.?ÿ You'll need to click on the Antenna tab at the bottom of the dialog box and choose Javad GNSS in the Manufacturer field, TRIUMPH-2 or TRIUMPH-LS (as appropriate) in the Type field, your height measurement method in the Method field (I always use Bottom of antenna mount), and the antenna height in the Height field.
6.?ÿ Click the OK button on the check-in dialog and you should have vectors ready to process.
7.?ÿ You'll probably want to add a coordinate to one or more of your points.?ÿ You can do this in the Properties dialog by clicking on the target (Add Coordinate) icon.
Done!?ÿ (It's easier than it sounds.)
Nice post!
Welcome to the site.
Only thing you were missing is some pics of the corners to show the kind of environment you took the shots in.?ÿ The aerial gives some idea, but not the whole story.
Thanks very much and appreciate for sharing the project with us.
Any experience with Javad LS on network RTK?
?ÿNorman Oklahoma What experience do you have with other modern GNSS setups? No doubt Javad works better than older GPS only rigs, but how does it stack up against modern competition in your experience?
My last experience was with the Trimble R8 model 2 (I believe...they went through some mods late in this model) this was the one with GPS / Glonass. When I started my own thing in 2014 I bought an R6 which was the guts of the R8 GNSS without the ability to receive the new constellations. I still use this with my one field crew on a daily basis on the NC RTN. Works like a champ in the open sky...been a great receiver. It's my understanding that the R10 works great and you can read about that on here...but the price tag is why I left Trimble...for GPS anyway.
Jim, thanks for the info...I tried this once...but I will try it again on this project and report to all if it gives me fixed baselines on this same site...seems like I recall it didn't on another site...more to come on that...
...I'll try and get some pix. I have to scratch for one more stone out there...or set the corner. There's three of these right up against trees that would give a good picture.
Nimal Any experience with Javad LS on network RTK?
I do have it configured to use either the NC RTN or Base/Rover via Verizon MiFi's. I would say that I do see a noticeable (but not really quantifiable) improvement with an onsite base. In the open...no...but in the tough wooded environments...Yes...onsite base definitely improves this scenario described above.
?ÿ
Thank you for the post and the reports. Based on the file path, I am glad I am not the only one using Google Drive as my company network.
?ÿ
Kudos.
Jim, are the precise GLONASS orbits at that link? Took a quick look, maybe I missed them?
SHG
I'm going to need to spend a little time with your post. More than I can at the office.....but you have spent a lot of time on it and it is appreciated just for that. A terrific way to introduce yourself.
I did not know that Carlson had a transformation routine like that.?ÿ I need one of those in my tool kit.. I have devised a way to trick StarNet into doing something like it....but that looks much more slick.?ÿ?ÿ
What experience do you have with other modern GNSS setups? No doubt Javad works better than older GPS only rigs, but how does it stack up against modern competition in your experience?
That transformation utility in Carlson is fantastic. It's a Helmert transformation (same as is used in a localization). It's an excellent tool for comparing two geometries simultaneously, such as survey to record (provided that the record is a modern, precise survey).
@TimmyG, this is very thorough. I've done quite a few tests like this with all sorts of types and brands of equipment over the years. It's a great way to demonstrate, particularly for yourself, what the capabilities and limitations of your equipment are. We should all be doing tests like this when we get new equipment (whatever type, whatever brand) so that we can prove the equipment and our competency with it.
Jim, are the precise GLONASS orbits at that link? Took a quick look, maybe I missed them?
Good question, and one that I've never considered before.?ÿ The files in the link above are described as "combination" ephemerides, which I assumed was GPS + GLN, but it looks like "combination" refers to a weighted analysis of GPS data from various analysis centers.?ÿ Accordingly, when I look into typical .igr and .igs files, all I see are GPS position and clock records.?ÿ
In order to get GLN orbits, it looks like you have to go here and get the .igv (ultra-rapid) or .igl (final) precise GLN orbits.?ÿ (There doesn't seem to be a GLN rapid product.)
?ÿ
Thanks Jim, that was my conclusion too, just wanted to make sure I hadn't missed something. I agree that GLONASS precise files seem to be a rare beast short of the final, which is why I often don't bother with anything other than broadcast.
Taking it a bit further, I have yet to see any flavor of Galileo orbits available, nor any of the other constellations. Doesn't affect me today as the hardware isn't there for us yet, however we will be upgrading one of the aircraft receivers to a multi-constellation unit soon and a number of CORS already have at least Galileo in the mix, we need more than broadcast orbits to use the data in a PPP solution, so staying tuned pretty closely to what is available and from where.
SHG
Timmy, did you shoot these points just once, or multiple times, and use the average of multiple observations?
I'll give you some of my observations. Certain times of the day, yield better coords than others. The dot scatter on the left hand side is a pretty good indicator of shot quality. If it shoots it fast, and has small dot scatter, it's usually pretty tight. If it struggles, string lines, and has large shot scatter, it's a poorer observation.
What's a total station, or a robot? (Wearing a Javad smile).?ÿ 🙂
Nate
