Static Control NetworkPosted by Williwaw on October 4, 2022 at 5:09 pm
I’ve got a challenging job mapping out a large subdivision that on the side of a mountain with lots and lots of big trees. Using RTK and 7-10 minute occupations, I brought in control for use with the total station, but just as I feared the precision just isn’t as good as I hoped. So I’m thinking of running a static network throughout the project area and leap frogging three receivers around for 45 minute overlapping observations with a fourth running continuously on my control out in the open a mile away. Is there any significant advantage to having a second continuously running receiver on another control point so that I effectively have two independent vectors running to each observed control point in the network? I’m thinking as an alternative I would just bring some of the closest CORS stations into the network, but they are some distance away. I’m hoping to significantly tighten things up.
- 19 Replies
- MemberOctober 4, 2022 at 5:40 pm
I have not had success with the 5 or 10 minute sessions, I hear are now possible with all of the GNSS constellations, either. Do what you said you want to do. When I do similar, I find little to no adjustable error at my control pairs or when I check into the next control pair. I believe it??s the multiple-simultaneous data that gives your post-processing software strong data. I typically use 30-minute sessions for control, keep two stationary and move a third. On larger jobs like yours, I keep two stationary, move two pairs around and finish the day with random additional spots. Good Luck.
- MemberOctober 4, 2022 at 5:57 pm
If you have access to another receiver, I’d vote for a second reference receiver 100%. Incidentally, that was the SOP for control work when I was on AK DOT&PF projects. At least it was for Central Region contracts. But we also had a lot of receivers at our disposal.
If you can’t get a second receiver I would run two separate sessions for the tree points, maybe 2×20 or 2×30 minutes.
I have found that having two sets of independent baselines really helped to tighten things up, especially if you’re in the trees to position the other points. If you have really large trees that are going to cut out pieces of your sky, it increases the likelihood that you’re going to have to edit out some cycle slips to improve the baseline processing. Full 4-constellation processing helps a lot there but I still get nervous if all I have is a single baseline to the primary control.
CORS might help, but unless things have changed since I left AK, most of them are GPS-only or at most GPS+GLO, which means that it is hit or miss as to whether baselines to points in the trees will fix at a reasonable precision. You’ll really need vectors between the CORS and those tree points to benefit from the CORS data, otherwise you’ll just be using the CORS to position the one or two reference receivers.
Not sure if you have obstruction sketches for the tree points but if you do some mission planning migth help. I do remember it being a challenge at those high latitudes when the SVs tended to be closer to the treeline than elsewhere.“…people will come to love their oppression, to adore the technologies that undo their capacities to think.” -Neil Postman
- MemberOctober 4, 2022 at 6:30 pmPosted by: @oldpacer
On larger jobs like yours, I keep two stationary, move two pairs around and finish the day with random additional spots.
I’ve used the method oldpacer describes in order to have that direct GPS vector between the two points that I would later be occupying with the total station. I would probably go with two stationary throughout the day and two working as a group to set point pairs. Some ties across point pairs can be helpful as well – move one unit forward for a half hour before bringing the second unit ahead.
Would be nice if possible to have the stationary ones separated by some distance (i.e. not both on the same site 1-mile away) IF at all possible.
- MemberOctober 4, 2022 at 6:56 pm
@rover83 Sound advice. Thanks. Being as you??re familiar with my area, this is Chugiak right up against Chugach State Park and the Chugach Mtns, which obliterates a significant part of the sky to the south and east, where constellations are this far north. Have four receivers, think I??ll go with two cooking stationary and leap frog the other two around. I??ll throw a sky plot together to identify the worst times for PDOP. In any event, sometimes RTK just has its limits and its time to do get down to simultaneous static and dust off my post processing skills.Willy
- MemberOctober 4, 2022 at 7:23 pm
It’s always interesting to survey against a mountain. I have two that climb over 2000′, I decided to place the surface more towards the lower elevations instead of a mean. State plane would really be short when the elevations are reaching 7000′.
- MemberOctober 4, 2022 at 7:39 pm
I’m skeptical about the value of running more than one stationary receiver unless you’re placing them far enough apart to see different SV arrangements or if you can’t get one in the open. It certainly won’t hurt anything, but if you have sketchy data at the rover it doesn’t matter how many fixed receivers you have, and if you have good data at the rover you’ll get good vectors from a single fixed unit that has a wide-open view. I just don’t see the payoff for the time spent running a second fixed receiver, not to mention the risk of it not being there when you go to pick it up.
- MemberOctober 4, 2022 at 8:52 pm
@jim-frame In not worried about receivers getting pinched, much less of an issue here than down there. Granted if rover data is too noisy, one or two base receivers becomes academic, however having two vectors some distance apart into the rovers would build more redundancy into the network over a single vector in the same fashion as tieing a point from two locations in a traverse would make for a stronger solution than a single side shot, which isn??t that far off from RTK. Mostly I just wanted to get others opinions on the subject, so, thanks for sharing.Willy
- MemberOctober 4, 2022 at 9:12 pm
It will be a trade off, setting the fourth receiver out for a base and running two or using the fourth to leap-frog. I would use it to leap-frog and save the time. How far away are the CORS. I spent years with the close ones being 130+ miles and 30-50 minutes will tie into them easily (usually), canopy will of course effect everything.
- MemberOctober 4, 2022 at 10:01 pm
Just how much area are we talking about? I’d be looking at traversing from point to point, then simultaneously adjusting the GPS vectors, which you already have, and the traverse data.
I doubt that you would get much better quality vectors by static than you have by RTK. Longer duration occupations do not make noisy data good. It just makes more noise.
- MemberOctober 4, 2022 at 10:14 pmPosted by: @williwaw
having two vectors some distance apart into the rovers would build more redundancy into the network over a single vector in the same fashion as tieing a point from two locations in a traverse would make for a stronger solution than a single side shot
This is the purported equivalence that I don’t buy. Tying a point from two terrestrial positions adds some strength to the new point, because the two ties are independent. But using a single tranche of rover data to compute vectors to two fixed receivers with the same sky view doesn’t actually add any redundancy, because the rover data is the same in both cases.
If the fixed receivers have compromised sky views, or they’re so far apart that they’re seeing even a slightly different set of SVs, then there’s some redundancy. Otherwise, I don’t think so.
I’m not trying to discourage anyone from running multiple fixed receivers, I’m just challenging the notion that it accomplishes anything useful.
Sometimes we do things just because they make us feel more comfortable, and that’s okay as long as the costs are understood. For me it’s way too many redundant control ties when doing urban topo — I just have a hard time not hitting a control point if I can see it, even though I know in my bones that doing so isn’t going to improve the topo. But I do it anyway.
- MemberOctober 5, 2022 at 2:03 am
@norman-oklahoma about a mile plus with a long wrap around spur and numerous cul de sac off shoots. Problem is there are numerous tight compound curves that with the terrain make avoiding short back sights impossible. Leaves are off and no snow yet so as good as it will get. As is, coming off RTK control, getting back sight checks exceeding a tenth or more. The only way I??ve managed to get them down to a few hundredths is by doing simultaneous static so many of the errors cancel out. Don??t get that with RTK. The reason on I??m looking at using two bases is simply given all the obstructions, I??m doubling the chances that one of the fixed receivers will have data that pairs with the rover??s data without being full of cycle slips, but if rover data is too noisy, doesn??t matter. Most likely I??ll be back next summer when everything is leafed out to stake the entire thing so now is my best chance to get the control situation to where I??m happy.Willy
- MemberOctober 5, 2022 at 12:40 pm
Can you share what brand of receivers you are using for this type of work? After using several different brands, I’ve only personally seen one setup that will do any good in canopy, and heard of maybe two others from reliable sources. Of course, all the salesmen have said for 20 years now, you can use this GPS in the woods….
- MemberOctober 5, 2022 at 1:39 pmPosted by: @jim-frame
This is the purported equivalence that I don’t buy.
Single vectors to a primary Control Point is a good and practical way to accomplish acceptable results. But like you said, I do it differently to make me feel more comfortable. I am old, duh, hence my alias; and much of how I do things is rooting in pre-GPS procedures. There is more to simultaneous-multiple observations than just two redundant vectors. You are forgetting the third vector created between the two base stations. This third vector acts as a backbone to strengthen your adjustment, will not get adjusted away very much and conveniently lies on your site for later use. I much prefer having a Control Baseline on site, rather than a Control Point and a Reference Point on site. Also, simultaneous-multiple observations transfer elliptical heights much better than two observations four hours apart. Most, actually all, of my work has a vertical component, so having a GNSS derived vertical that exceeds vertical accuracy requirement is necessary. My rover is control and my bases are autonomous, so I do not believe vectors to two separate points to be redundant. As far as time and cost, I believe them to both be less, but as slow as I am, it??s a moot point. Safety, yes, it??s hard enough to find one good safe spot on the site, much less two. And yes, I have asked the lady sweeping her porch if I can place this in her yard while I am working.
- MemberOctober 5, 2022 at 2:08 pm
@learner Trimble GNSS R83 and R10, full constellation. But to clarify using receivers in more relatively open areas to establish control to traverse off of with S6. If I have hundreds if not thousands of shots to pick up under canopy in canyons, gun is way to go. Just too much slop with RTK in those conditions and is painfully slow.Willy
- MemberOctober 5, 2022 at 3:00 pm
Thanks! Good luck!
- MemberOctober 5, 2022 at 7:03 pmPosted by: @jim-frame
unless you’re placing them far enough apart to see different SV arrangements or if you can’t get one in the open.
That’s often the case for me. I can easily find fields that will provide some decent sky view, but not always completely open. With receivers in different fields with (even slightly) different sky view at the stationary receivers, I can pick and choose the most “open” areas of a site to set point pairs in. As I move around a site, the sky view at the moving units may shift from being similar to one or the other stationary receivers. Most of the time, there is a good solution from both stationary receivers, but if one fails to get a good fixed solution, it is unusual that both fail.
When pushing things in the open areas of the woods, I also just like being able to have two different solutions that indicate the same point. So, I’ll process in groups that use one then the other stationary receiver to compare. Of course, I’m an old fogey that hasn’t yet gotten the comfort level of just running out in the woods with a GPS unit and feeling good about the reported coordinates.
- MemberOctober 5, 2022 at 10:00 pmPosted by: @jon-payne
I can easily find fields that will provide some decent sky view, but not always completely open. With receivers in different fields with (even slightly) different sky view at the stationary receivers, I can pick and choose the most “open” areas of a site to set point pairs in. As I move around a site, the sky view at the moving units may shift from being similar to one or the other stationary receivers. Most of the time, there is a good solution from both stationary receivers, but if one fails to get a good fixed solution, it is unusual that both fail.
That’s been my experience as well. It’s rare that conditions are exactly the same at both reference receivers. Ultimately, I’m bringing CORS (NGS or state/local) data into my project and processing as many of the baselines as possible.
What I usually end up with is all of the CORS-reference vectors processing well, about 25-50% of the CORS-woods vectors processing well after cleaning up cycle slips, and 80-100% of the reference-woods vectors processing well after cleanup. All depends on how gnarly the canopy is of course.
If I have four CORS (pretty standard), two reference receivers, and ten woods points, assuming things process as above, a single observation session will give me about 35-45 vectors against the 12 unknown points (assuming CORS are all fixed in adjustment). This is single-baseline processing leapfrogging, not multi-baseline session processing, so the only vectors I am removing are the ones between the CORS.
Would it be more rigorous to run two different sessions on two different days? I would say yes.
Can I hit desired specs while being more efficient by running dual-reference on a single day? I would also say yes. It all depends…“…people will come to love their oppression, to adore the technologies that undo their capacities to think.” -Neil Postman
- MemberOctober 6, 2022 at 1:17 am
My jobsites tend to look like this:
Log in to reply.