NavList:

   

Reply

We often come back to this, but of course technology moves on. We
discussed using digital camera and using the pixels to measure the
angles between the sun and horizon. Of course one could have two
cameras at a calibrated angle so the Horizon and sun are in the field
of view of each, or actually make a sextant and use one camera (but
cameras are cheap and the best one for the horizon would not be best
for the sun). My thought is that digital image stabilization has
improved on cameras (perhaps thanks to smart phones), and stabilized
platforms for video cameras are also available at a consumer level now
(for accounting for larger motions where the object of interest would
go out of view).

If such a system was mounted on a motorized three axis stage and could
average over a lot of snapshots how close would we be to a "consumer"
cost fully automated sextant that could deliver a line of position
with error bars as long as it could see the Sun and Horizon?

Of course in the spirit of Frank's GPS anti spoof app, it could
compare with the GNSS as well as dead reckoning information. Dare I
say (with the current meeting in progress at Bletchley) it could
"learn" to take better sights, eg timing with waves, upper limb when
lower limb obscured, judging position of horizon, so if GNSS failed it
would work better based on "practice".  One might of course call this
"continuous calibration" but the fact is the "Machine Learning"
enthusiasts have developed some really fast general purpose
optimization code for such calibration.

Is there anything other technology on the horizon (sorry!),such as
improved low cost accelerometers, that would make this more viable?
Could we do without the horizon on a ship or yacht if they were good
enough?

Sorry I meant to add references to previous NavList discussions but I
have to get back to work, will try to look them up later.

Best wishes

Bill Lionheart

   

Reply