There are a bunch of simplifying facts...

1) The favorite star is Canopus, as I understand (I'm no astronomer) it's
the brightest, it's close to the ecliptic, and there are no other bright
stars nearby to be confusing.

2) You know pretty nearly where you are by your inertial system,  you're
merely trying to improve on it, so you can point your telescope pretty
close.

3) The image is a single bright point.  Your "retina" only needs 4 cells.
You move "up" or "down" until the "up" pair of cells matches the "down" pair
of cells, and similarly for the "left" and "right" ones.  You confirm that
you're pointing to the right star by checking the absolute brightness of
what you see.  If it's not bright enough, it's the wrong star, you were
looking for the brightest one.

To get aligned in 3 dimensions, you need one other star, preferably about 90
degrees away.  I think Polaris is the favorite for that, the familiar north
star.

4) An inertial navigation system intrinsically gives your location
"inertially", that is to say, it doesn't turn with the earth, so it doesn't
move around the sky with time of day or seasons, as long as the initial
position that was set just before launch was chosen that way.

For a weapon system that could be launched at any time, I guess it must be
able to use an alternate star in case the one it wants happens to be on the
other side of the earth at the time.

Star tracking should be able to null out the gyro drift, at least during the
coast phase, though I can't see how it could help for accelerometer errors.

> -----Original Message-----
> From: M. Adam Davis [mailto:adavis@UBASICS.COM]
> Sent: Thursday, February 08, 2001 11:19 AM
> To: PICLIST@MITVMA.MIT.EDU
> Subject: Re: [OT]: How to knock a missile out of the sky...?
>
>
> Well, since they are above the atmosphere (or at least in a
> region where
> the visibility is a constant), then thye don't have to do
> nearly as much
> processing as one would think, but even then I've got to ahdn
> it to the
> people who make it work.
>
> Point a camera away from the earth - there will be several
> stars which are
> brighter (and maybe color variations are used) than the rest.
>
> At that point you can reference your real time clock (which
> is accurate)
> compute the positions of the stars, match up the positions to
> get a 3d fix
> on the camera, and then compute the earth's position and
> rotation (based
> on the RTC), find the error in the inertial navigation and
> correct for it.
>
> But image processing is probably the hardest part for me to
> conceptualize.  Since the image is simply a black background
> with white
> dots it should be significantly easier than, say, robot
> vision on earth.
>
> -Adam
>
> "Sean H. Breheny" wrote:
> >
> > Can anyone point me to more info on how they do star
> sighting? That seems
> > to me to be a fairly difficult problem. It would be
> especially interesting
> > to see how they solved it with 1950s or 60s electronics.
> >
> > Sean
> >
> > At 08:54 AM 2/8/01 -0800, you wrote:
> > > >    Missiles that are only going to be fired on
> cloudless nights can
> > > >    take star sightings.
> > >
> > >
> > >They take the star sighting after they leave the
> atmosphere to correct for
> > >errors in their inertial nav.
> > >
> > >John
> > >
> > >--
> > >http://www.piclist.com hint: The PICList is archived three
> different
> > >ways.  See http://www.piclist.com/#archives for details.
> >
> > --
> > http://www.piclist.com hint: The PICList is archived three different
> > ways.  See http://www.piclist.com/#archives for details.
>
> --
> http://www.piclist.com hint: The PICList is archived three different
> ways.  See http://www.piclist.com/#archives for details.
>
>

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.