WAAS (SBAS) accuracy

At one time several years ago I would setup my construction jobs using a N and E shift so that the lat/long in the data collector was WGS84 but the coordinates were NAD83. I did this so that rough staking or stationing with waas would be a bit more accurate. Then we started using the TxDot rtn on several jobs so I quit cheating. Problem is, I'm the one with the 10 year old DC that does not have internet access.

I didn't do any type of testing like you did but I found that the vast majority of the time I was less than a meter horizontal.

James

Being a GIS guy, many in our industry are sold on WAAS accuracy - witness the plethora of bluetooth WAAS receivers on the market.?ÿ I have done hundreds of tests since WAAS inception, and being an Alaskan, our benefit up here are in many cases, no trees, open area and many Ground Reference Stations that build the real-time IONO maps necessary for WAAS high quality augmentation.?ÿ A high end Trimble mapping receiver using best practices are typically well under 50 cm horizontal.?ÿ Shifting to NAD83 is paramount.?ÿ Much of the early WAAS testing did not compensate for this shift as discussed by John Hamilton.?ÿ Some receivers like to "settle", while others like EOS and GENEQ receivers coast with additional algorithms that can coast into areas with obstructions, but they run the risk of using stale IONO predictions.

?ÿ

Joel

A typical high end Trimble mapper like Joel describes, using best practices, should be able to achieve about 30cm Hz, at least in theory.

Hi John,

I believe that WAAS is IGS08 current epoch, adjusted on the 1st of each year. At one time I think that I had an official document that expressed this, but I can no longer find it. Do you have anything that definitively describes the frame and epoch? I would like to archive it.

A couple weeks ago I ran an SP20 for 4.5 hours on the robot, looking at range of measurements. So dump, then 30-second wait, then store a single epoch. Ran under moderate canopy (because that is how I use them typically). This is a different test than letting a receiver log data without a dump for the same period of time. It forces a re-acquisition between every measurement. I think that it mimics real world use where one is moving around under canopy, setting the receiver down and acquiring solutions from scratch between measurements. I will include a video and sample result below.

I think that when we talk about DGPS/WAAS we need to make it clear that we are looking at CEP which can be a lot more optimistic than we state for FIXED measurements. In other words: 1-sigma might be 0.26 meters, but the range of measurements might be a couple meters.?ÿ

Here is a video:

and here is a summary of the results:

The 2D 1-sigma (to the average) is 0.26m, but the range is over 3 meters. Kurtosis can be statistically disappointing...

?

I think the demise of the USCG (and USACE) beacons can be attributed somewhat to WAAS.?ÿ

I posted this the other day, but here it is again...for those that don't know the FAA has established a reference station network across the US, usually at ARTC (ATC) centers, but also a bunch in Mexico and Alaska. Here is a list:?ÿWAAS Ground Segment

and a map:

The station at Mexico City airport is sinking (along with the rest of Mexico City):

?ÿ

Mark,

Without the benefit of that awesome testing machine, my tests are done with students in my 3-day GPS course when on the third day, they mount their Trimble device (or antenna mount) on a pole set over an OPUS control point.?ÿ They shoot in 30 points at 5 epochs each (logging rate 1/pps) and compare their coordinates after controlling for datum in an ArcGIS environment.?ÿ I do think much is lost when conducting accuracy tests when in real - life that feature is laid to rest in a GIS and stored as "accurate".?ÿ Datum demons creep in, and some software (ArcPad was one of them) had 5 datum gates to pass thru before storage.?ÿ One mistake on a gate and accuracy is wasted.?ÿ So much of the datum nightmare that I see in GIS is what the software does and does not do properly even if we are limited to 7 parameters.?ÿ In our recent 3-day AutoCad Map3D course I was shocked by the lack of defined NAD83 definitions in the most current version of AutoCad Map 3D version 2019!.?ÿ It appears to me much is being lost when after a CAD dataset is held to OPUS, a coordinate system Alaska State Plane has to be still defined as "NAD83".?ÿ Software manufacturers are not helping.

?ÿ

Joel

Mark...excellent tool you have developed for testing....

Joel, you are preaching to the choir brother!

I wrote an article titled 'There is no Geodesy at ESRI' about 4 years ago but nobody would think about publishing it. One editor said "don't ever send me anything like that again!" Evidently it is not okay to say anything, even slightly negative, about any current or potential advertiser. (Wendell, I will totally understand if you delete this comment.)

I spent an entire day trying to get a decent GEOID inserted into test data that I stored in Collector a few months ago. The receiver was logged into a network and the position was better than 1 cm (XY&Z) but I was getting EGM translated orthometric heights stored. I finally just fixed them in a spreadsheet. Collector should back out the transmitted EGM GEOID height to get ellipsoid, then subtract off the Geoid12B. I am sure that I was missing something, it should not be that hard. I found some post collection workflows online, but honestly I could not understand the instructions well enough to implement them. And that certainly will keep you from using Collector to stake.

I have been playing around with my new SP20 receiver (which is $@%# amazing) and I am preparing for this battle again. It is disappointing to blow a 1 cm vertical measurement by using a GEOID model that has a 6 cm bust to G12B.

Back on track...

Do you have any good documentation the frame of WAAS?

When you are looking at WAAS, you need to look at very-very long data sets. I think 24-hour sets may be a bit short, 48-hour sets probably are about right. And I am not sure if it is good enough to just let a receiver cook because I get totally different answers when I dump.?ÿ

?ÿ

JohnH: You are welcome to run receivers on Mr. Robot anytime. It is shocking the stuff that one can learn after a few thousand dump cycles. Shocking! The best thing about Mr Robot is the stats and graphs are produced in real time. So when the test is complete you don't have to spend 3 hours screwing around processing data. The results are just sitting on a Google Drive, filed by Year Month Day and test number.

?ÿ

Here is a document I found...

https://www.gps.gov/technical/ps/2008-WAAS-performance-standard.pdf

?ÿ

But NOWHERE does it mention ITRF or WGS or NAD or reference frame or...

Posted by: Mark Silver

Joel, you are preaching to the choir brother!

I wrote an article titled 'There is no Geodesy at ESRI' about 4 years ago but nobody would think about publishing it. One editor said "don't ever send me anything like that again!"

?ÿ

Mark, it is not politically correct to talk about ESRI, they invented "GIS" and know all! They have done a great disservice by faking so much geodesy, a lot of their so called "re-projections" do nothing more than change the label so the metadata "looks correct". Big pet peeve is to hear mapping folks in my own company say we will "re-project" the data using ESRI products or Blue Marble, etc. No one has a clue what is or isn't going on! A lot of this stems from the original WGS84 and NAD83 being approximately the same (was the goal anyway) and as those have become more refined, the software continues to blur the distinction, a shame really. BTW, ESRI isn't the only software still treating them the same, but a huge market leader with a huge following many of who don't know what they don't know.

SHG

I hear it all the time...we need the coordinates on WGS84...

Mark, it is not politically correct to talk about ESRI

Oh there you go again...

They are in a position where they could single-handedly teach all of the GIS professionals about:

1. Survey Feet vs. International Feet

2. Orthometric vs. Ellipsoid elevations

3. Reference Frames and measurement epochs

And since they are hell bent on storing coordinates, they could do it in a way that would stand the test of time. So that in 200 years you could automatically convert a today coordinate to a future coordinate and it would be right.

They control all the course content and it really would make the world a better place, which I think they are all about.

If I die early, it will probably be from exasperation from dealing with new GIS and new UAV people and companies. ?ÿ

In defense of ESRI, their projection team including Melita Kennedy is doing the best they can do without time dependent shifters.?ÿ The bookkeeping transformations which Shelby Griggs refers too (NAD83=WGS84) is prevalent in all software - the names may be different, but the 0,0,0 3-parameter "shifter" is a must, and dare I say, thee most common shifter in any workflow that demands "doing nothing" on output.?ÿ If GIS was not offered the ability to transform then your stuck with Google Earth.?ÿ ?ÿCollector (don't get me started) is maturing, but its extremely mis-leading to those without training, a point highlighted by Mark Silver's attempt above to keep collector out of the way blowing a 1 cm incoming stream.?ÿ The whole basis of Collector is using, out of the box Web Mercator (WGS84) as the basis for computation - all in the name of making something user-friendly, world-wide and efficient for slicing up the entire earth in a fast-easy-to-display environment.?ÿ GIS specialists have their work cut out for them to make the right decision, and in my opinion, were again, losing valuable trust that was being built between our industries.

Diligent GISer's using good equipment, proper adherence to CORS and holding those coordinates are doing the best they can but still cannot control for time dependency (exception is Blue Marble).?ÿ The projection engine inside ESRI is excellent, and has long left the camp of NAD83=WGS84.?ÿ Exposure of the parameters can be easily done and compared to SNAY/SOLER articles outlining parameters or in the most current reference frame, compare with NGS DATUM page and confirm without a doubt, with the exception (again) of time dependency, lock onto the exact same parameters.?ÿ While we beat up ESRI, AutoDesk, Pix4D, Agisoft, Trimble are all in a mess when it comes to datums and allowing a competent GIS specialist to align data from various sources.?ÿ ?ÿHeck, AutoCad Map 3D 2019 has no ability to assign NAD83 (2011) to a Drawing file.?ÿ How insane is that??ÿ PIX4D calls WGS84 a coordinate system!?ÿ?ÿ https://support.pix4d.com/hc/en-us/articles/202558239-Select-Image-GCP-Output-Coordinate-System-?ÿ?ÿ

Wishing all of us could hold a VRS with dual frequency GPS receivers, but thats not in the cards.?ÿ We did build the foundation for those GIS devices in cars though.....

Joel

Mark?ÿ - I asked Eric Gakstetter to chime in on the current reference frame of WAAS.?ÿ He knows.

WAAS INFO.?ÿ ?ÿFrom Erik Gakstatter