Notifications
Clear all

How to evaluate uncertainty in an azimuth

57 Posts
12 Users
0 Reactions
9 Views
(@mathteacher)
Posts: 2081
Registered
 

@john-nolton

I'm sure it will.  Of course, I said surveying standards, not geodetic standards, but I get your point.

 
Posted : 24/07/2020 3:36 pm
(@mathteacher)
Posts: 2081
Registered
 

@cf-67

You gotta be careful about them generalities. See below.

 
Posted : 24/07/2020 3:40 pm
(@larry-scott)
Posts: 1049
Registered
 

@cf-67

i get rinex online, from others, I’ve had zero issues. My Ashtechs are weekrollover glitched. So I generate rinex, fix the date and leave it in rinex, and GNSS likes it fine.

And in your task L1 only will still yield best az

maybe it’s the version of rinex that you’re generating. Most can select 2.11, and think that’s the most recognized. (I’ve not tested rinex 3) If GNSS won’t take your rinex, does opus like it?

you might post this on the GNSS page

 
Posted : 24/07/2020 3:53 pm
(@cf-67)
Posts: 363
Registered
Topic starter
 

@larry-scott

Ok, I'll have another go and if no luck I'll start a new thread in GNSS. 

 
Posted : 24/07/2020 4:13 pm
(@larry-scott)
Posts: 1049
Registered
 

@cf-67

and a working GNSS baseline processor is good thing to keep around. It’s not a Win10 issue

 
Posted : 24/07/2020 4:24 pm
(@a-harris)
Posts: 8761
 

The system looks to be controlled by GPS.

How do you calibrate GPS?

 
Posted : 25/07/2020 12:35 am
(@bill93)
Posts: 9834
 
Posted by: @larry-scott

@cf-67

If GNSS wonƒ??t take your rinex, does opus like it?

We had a thread recently where a file wasn't being accepted and it turned out the way it was generated under Win10 it used 16-bit characters. Changing the process to get 8-bit characters worked.

https://surveyorconnect.com/community/software-cad-mapping/rinex-file-will-not-import-to-tbc-5-20/

 
Posted : 25/07/2020 4:41 am
(@bill93)
Posts: 9834
 
Posted by: @a-harris

The system looks to be controlled by GPS.

How do you calibrate GPS?

It is likely the heading sensor mentioned is in addition to GPS and at a minimum needs to be rotated into agreement with the boat heading, and possibly also corrected at various azimuths.

GNSS could perhaps be used to get heading while in straight-line motion, but probably not to the precision mentioned.

 
Posted : 25/07/2020 4:51 am
(@john-putnam)
Posts: 2150
Customer
 

Correct me if I'm wrong, but this system appears to use multiple GNSS receivers, pulse and IMU, to determine azimuth.?ÿ The GNSS derived baseline azimuth will be very precise at 1 mm error over the 4 m baseline.?ÿ The IMU appears to me mostly used for pith and roll, notice the drift in azimuth when GNSS is down.?ÿ In theory, you could add more receivers and do away with the IMU all together.?ÿ I think it is more critical to obtain the relationship of the sensors baseline to the axis of the vessel.?ÿ Just like surveying in an aerial mapping plane for offsets from the sensors to the camera.

 
Posted : 25/07/2020 6:33 am
(@ashton)
Posts: 562
Registered
 

I'm not sure how to do this analytically, so I tried a Monte Carlo approach. I interpreted cf.67's post to mean the distance from the true point to the measurement was normally distributed, with 2-sigma being 10 mm, so 1 sigma is 5 mm. The angle?ÿ is random and uniformly distributed from 0 to 2 pi radians. I'm using the Cartesian coordinate system where 0 degrees is the positive x axis. Let there be two points, the first at (100 m, 100 m) and the second at (100 m, 200 m).

For each trial, I generate a random offset from the first point, and a different random offset from the second point. The I find the azimuth from the first to second point, which is nominally 0 degrees.

In Excel, I created 1000 trials. The angle was +- 20 arcseconds (2 standard deviations)

In an earlier post it was stated the sideways 2-standard-deviation would be 14 mm. Applying this to a 100 m distance, the angle would be +- 29 arcseconds.

 
Posted : 25/07/2020 11:26 am
(@cf-67)
Posts: 363
Registered
Topic starter
 
Posted by: @john-putnam

I think it is more critical to obtain the relationship of the sensors baseline to the axis of the vessel

Yes, that is definitely one thing the client has said is required.

 
Posted : 25/07/2020 12:00 pm
(@cf-67)
Posts: 363
Registered
Topic starter
 

@ashton

Thanks for taking the time.

I thought maybe the "propagation of errors"  formula could be arranged to evaluate how coordinate uncertainty translates into azimuth uncertainty, but I was overthinking I think.

Without baseline processing software, I knew single point positioning of each point would only get me +/- 10mm. However I seem to have managed to overcome the difficulties I was having with GNSS Solutions and will now use that.

 
Posted : 25/07/2020 12:16 pm
(@larry-scott)
Posts: 1049
Registered
 

@cf-67

What was the GNSS hang up? Just curious. (It’s my only baseline processor at present.)

and it should be easy enough to make 2 observations, even at the obscene 1 hr ea. 

 
Posted : 25/07/2020 1:20 pm
(@larry-scott)
Posts: 1049
Registered
 

@bill93

i make no accommodation for 64 bit win10. I loaded it and it worked right off. I may not rennet the ins and outs right off, but I never had to abort. 

 
Posted : 25/07/2020 1:25 pm
(@richard-imrie)
Posts: 2207
Registered
 

I'm going to step onto the nautical gang plank here and say FWIW, that heading is the direction vessel is pointing, and that's not necessarily the direction the vessel is going.

 
Posted : 25/07/2020 2:10 pm
Page 3 / 4