Is this a fair comment?
While it takes at least a few hours of static to arrive at an accurate position for a point, the azimuth/orientation would appear to be nailed down to survey grade accuracy immediately.
My reasoning...
If you were to undertake a topo in assumed or "read from GPS" coords, while collecting static on the base to submit to OPUS or PPP - then translated the job to that, (and this is a standard procedure and common use of OPUS or PPP) you would be refining the positions but you wouldn't be refining the orientation any - and it doesn't seem to require it.
Apparent proof...
If I set up my base on a known UTM point, then drive several kilometres to any number of other known points in my city, I get a good check (maybe 10-15mm or less). I have only given the base a known point to play with - no 2nd point or azimuth required.
Is it just a result of the considerable redundancy and geometry of the system that azimuth is so quickly determined?
If You Are In A Hurry, Do Solar Azimuths
It takes 20 minutes to get a survey grade position via OPUS-RS.
In 20 minutes one can get survey grade point to point relative vectors and azimuths.
The several hours you talk about ate to get a survey grade position irrelevant of every other observation you do. A surveyor cannot make a living doing a single observation in several hours, so please redefine your stated question and purpose.
Paul in PA
If You Are In A Hurry, Do Solar Azimuths
I'm not in a hurry - it was an observation, that in an RTK environment with only a base station coordinate, the accurate orientation seems to be resolved immediately, something borne out by later checks. I am interested in the science behind that.
The Science Behind That Is That....
The constellation of GPS satellites is rather accurately placed in the sky relative to each other horizontally via broadcast ephemeris.
Raw elevation can still be up to 100' off.
The best thing you can do on immediate post processing is to place a point at a good elevation. That brings the relative positions of the lot into great conformity with the ultimate solutions.
Paul in PA
Not real sure but I think it may boil down to how far away your Az Ref point "backsight" is.
ex. you can be a foot off in Lat-Lon for your Az Ref point
if your Az Ref point is 40 miles away and
still have one second of accuracy.
But a foot off in Lat-Lon for two points 1000 ft apart = big problem
> ex. you can be a foot off in Lat-Lon for your Az Ref point
> if your Az Ref point is 40 miles away and
> still have one second of accuracy.
>
Then if you have multiple "backsights" (the satellites), 22000km away, that might explain the accuracy...
Good point!
(I realize GPS measures distances to satellites, not angles, so they are not really backsights)
I think it may be more related to the ground Ref Sta you are using.
Say for example the CORS station(s) your GPS is being corrected to.
Say your GPS is reliably "accurate" to 0.2 ft
It will be that way just about anywhere on Earth
If you were to use two points located to that accuracy and
they were 1000 ft apart,
The Inv Az between them would NOT be very reliable.
But if those two points were 10 miles apart
(still only "accurate" to 0.2 ft)
The Inv Az between them would be very reliable
> Is this a fair comment?
>
> While it takes at least a few hours of static to arrive at an accurate position for a point, the azimuth/orientation would appear to be nailed down to survey grade accuracy immediately.
>
>
hello Colin,
yes, as i understand it, orientation/azimuth is 'built-in' to GNSS. i don't see how it could work otherwise.
GPS derives its azimuth and scale from the coordinates of the tracking stations. Since these stations are known to sub cm accuracy, and are separated by 1000's of km, the scale and azimuth accuracy of the system is VERY good. You can get very accurate scale and azimuth with no known position, i.e. by processing a baseline standalone. Since a typical C/A code position is 5 m or better, this would introduce an insignificant error into the results. In the "old" days of GPS, when SA could contaminate the results by many 10's of meters (up to 100), it was important to have a tie to a known station to eliminate this bias.
BUT, you still must deal with the uncertainty at each end of a baseline. If you have 2 cm accuracy at each end of a 500 meter line, then potentially you could be off 4 cm, or 16" of arc. 5 mm at each end would be 4" worst case. Typically the azimuths are better that the worst case, of course. One must also worry about centering, multipath, etc.
When I first started doing GPS (mid 80's), it was very time consuming and expensive to do azimuth pairs using GPS. We would typically set a single point, and use a T-2 or T-3 to observe polaris to determine the azimuth to the azimuth mark. I did a whole county (80 points and 80 azimuth marks) that was bluebooked. It got to where I could do 5 or more second order azimuths in a good night (and still had a day job!) using a T-3.
> GPS derives its azimuth and scale from the coordinates of the tracking stations.
That's neat! GPS is amazing - thanks for the info
Colin,
I’m not sure if this is along the lines of what your are asking but it has been my experience that the azimuth I get with setting up a random base station and then doing rtk on a project closely matches the final azimuth I derive after opus processing. Due to the fact that some of my work is remote and far from a source to submit opus data we regularly set up a good base location let the unit decide where it’s at and go with GPS azimuth to conduct the survey. We then do static on a min of three points and when we return to civilization we submit to opus. Our results have been pretty amazing as far as the difference between raw azimuth and opus derived azimuth. Most times it isn’t really worth the rotation.
We used to do solars but the sun isn’t very reliable around here.
yes, exactly along the lines I'm getting at - that the azimuth determination is spot on almost immediately - and that's pretty clever.