Star*Net version 6 pro and Leica GPS vectors

ekillo, post: 368134, member: 773 wrote: I know that I could fix the bearing from my base to the farthest point according to the GPS inverse between those points and would be good to go on this project and manually compare the other GPS inverses, but wanted to know the best way for establishing a future work flow of including the GPS vectors into the project. I did not know if you used an inline command or can Star*Net read the GPS (Leica .moo) file?

Do you have the option of exporting the Leica vectors in ASCII format? You're asking a couple of questions, but the first step is to figure out whether you can import Leica GPS vectors into Star*Net V.6-Pro. The user manual indicates that Leica vectors need to be exported from their database to an ASCII format that can be read.

Kent McMillan, post: 368140, member: 3 wrote: Do you have the option of exporting the Leica vectors in ASCII format? You're asking a couple of questions, but the first step is to figure out whether you can import Leica GPS vectors into Star*Net V.6-Pro. The user manual indicates that Leica vectors need to be exported from their database to an ASCII format that can be read.

LGO will probably do that, I have never tried.

Thanks,
Ed

Since this question is related to Star*NET and I have been exploring this software lately, when I was reading Ghilani's book it says: "When more than minimal control is held fixed in an adjustment, the observations are forced to fit this control. For example, if the coordinates of two control stations are held fixed but their actual positions are not in agreement with the values given by their coordinates, the observations will be adjusted to match the erroneous coordinates. Simply stated, precise observations may be forced to fit less precise control".

Is there a way to put weight on the controls when using Starnet?

GeomÌÁgico, post: 368144, member: 11554 wrote: Is there a way to put weight on the controls when using Starnet?

Yes, the coordinates that are entered as conditions of the adjustment may be assigned standard errors for their components. They don't have to be fixed. However, unless one knows something about the uncertainties of the coordinate values entered, in relation to the coordinate held fixed, just guessing at them will tend not to be very satisfactory.

Kent McMillan, post: 368145, member: 3 wrote: Yes, the coordinates that are entered as conditions of the adjustment may be assigned standard errors for their components. They don't have to be fixed. However, unless one knows something about the uncertainties of the coordinate values entered, in relation to the coordinate held fixed, just guessing at them will tend not to be very satisfactory.

Ghliani says "It is possible to weight a control station according to the precision of its coordinates. Unfortunately, control stations are published with distance precisions rather than the covariance matrix elements that are required for weighting. However, estimates of the standard deviations of the coordinates can be computed from published distance precisions. That is, if the distance precision between stations A and C is 1 : 10,000 or better, their coordinates should have estimated errors that yield a distance precision of 1 : 10,000 between the stations."...."From the coordinates of A and C, distance AC is 2542.65 ft. To get a distance precision of 1 : 10,000, a maximum distance error of å±0.25 ft could exist. Assuming equal coordinate errors yields"

In this example A and C are the controls. I am not sure why, but when I tried it using Starnet, it said: *WARNING Network Has No Fixed XY Stations*
any idea why?

GeomÌÁgico, post: 368146, member: 11554 wrote: In this example A and C are the controls. I am not sure why, but when I tried it using Starnet, it said: *WARNING Network Has No Fixed XY Stations*
any idea why?

You need to fix at least one of the XY stations and the uncertainties of the other coordinates should be represented as standard errors with respect to the fixed station. If for some reason you want to pursue a weighting scheme based upon an uncertainty expressed as a ratio of the distance, then you can simply enter the bearing ("B" data type) and distance ("D" data type) from the fixed station to the other control point with standard errors of bearing and distance that correspond to those for the separation.

Kent McMillan, post: 368147, member: 3 wrote: You need to fix at least one of the XY stations and the uncertainties of the other coordinates should be represented as standard errors with respect to the fixed station. If for some reason you want to pursue a weighting scheme based upon an uncertainty expressed as a ratio of the distance, then you can simply enter the bearing ("B" data type) and distance ("D" data type) from the fixed station to the other control point with standard errors of bearing and distance that correspond to those for the separation.

Not sure if I understood it 100%, this is the code I have so far, where:
1 denotes A
2 denotes B
3 denotes C
4 denotes D
5 denotes E
6 denotes F
->
#Coordinates
C 1 10000.00 10000.00 ! !
C 3 10528.65 12487.08 ! !
C 2 11103.000 10862.000
C 4 9387.000 11990.000
C 5 9461.000 10948.000
C 6 10131.000 11595.000
# Distances
D 1-2 1400.91 0.023
D 1-5 1090.55 0.022
D 2-3 1723.45 0.023
D 3-6 976.26 0.023
D 3-4 1244.40 0.023
D 2-5 1644.29 0.023
D 2-6 1217.54 0.022
D 4-6 842.75 0.022
D 4-5 1044.99 0.022
D 5-6 930.93 0.022

so you are saying I should add "D 1-3 2542.65 0.25" ? and let point C free? I tried it but then I dont have any orientation constraint.

I haven't tried it in Star*Net, but there is no mathematical reason you have to fix any coordinates. My own program runs with standard errors on all coordinate inputs as well as other measurements - nothing fixed.

Maybe it's the routine that finds starting values that has trouble? I supply initial guesses for every point to my program because I haven't written an algorithm to come up with them.

Bill93, post: 368167, member: 87 wrote: I haven't tried it in Star*Net, but there is no mathematical reason you have to fix any coordinates. My own program runs with standard errors on all coordinate inputs as well as other measurements - nothing fixed.

Maybe it's the routine that finds starting values that has trouble? I supply initial guesses for every point to my program because I haven't written an algorithm to come up with them.

I understand, I could do it in Matlab as well but for larger projects would be more time consuming.

GeomÌÁgico, post: 368153, member: 11554 wrote: Not sure if I understood it 100%, this is the code I have so far, where:
1 denotes A
2 denotes B
3 denotes C
4 denotes D
5 denotes E
6 denotes F
->
#Coordinates
C 1 10000.00 10000.00 ! !
C 3 10528.65 12487.08 ! !
C 2 11103.000 10862.000
C 4 9387.000 11990.000
C 5 9461.000 10948.000
C 6 10131.000 11595.000

If you wanted to assign, say standard errors of 0.01 and 0.01 to the coordinates of 1 and 3 and let the others be free, here is what the data entry would look like. This runs fine, BTW.

C 1 10000.00 10000.00 0.01 0.01
C 3 10528.65 12487.08 0.01 0.01
C 2 11103.000 10862.000 * *
C 4 9387.000 11990.000 * *
C 5 9461.000 10948.000 * *
C 6 10131.000 11595.000 * *

Kent McMillan, post: 368179, member: 3 wrote: If you wanted to assign, say standard errors of 0.01 and 0.01 to the coordinates of 1 and 3 and let the others be free, here is what the data entry would look like. This runs fine, BTW.

C 1 10000.00 10000.00 0.01 0.01
C 3 10528.65 12487.08 0.01 0.01
C 2 11103.000 10862.000 * *
C 4 9387.000 11990.000 * *
C 5 9461.000 10948.000 * *
C 6 10131.000 11595.000 * *

Thanks a lot Kent! It worked perfectly!

Cheers!