Re: inertial systems

At 11:52 PM 5/30/2004 -0700, you wrote:
>On Sun, 2004-05-30 at 12:04, John Carmack wrote:
> > Was the code I gave Pierce any help in getting the 3D inertial
> > transformations working with your crossbow?  For really short flights with
> > no significant wobble you might get by with just integrating rates, but
> > eventually you will need 3D transformations.
>
>Yes, although I've gotten a lot more from looking at the functions that
>get data from the IMU and do the screen drawing. Your code has been
>extremely helpful all the way around.

That is from my "quick hacks" directory, so don't read too much into it...


>I've never done graphics programming before, so looking at your code and
>reading the OpenGL manual pages has got me on the right track for that.
>I hope to have the display done tomorrow.

The biggest hurdle for starting 3D graphics programming is getting anything 
at all drawing on the screen.  It is really easy to just have the 
projection or model matrix putting things behind you, or have the triangles 
drawn backwards so they are culled, or have some critical functionality 
disable or enabled with the wrong parameters.  Modifying sample code that 
puts something on the screen is a good way to star.

>  That would allow us to move
>Gizmo around and see similar motion on the computer display. Once we're
>happy with that I'll add in the PID loops and try controlled flight.
>
>Matrix math has been the bane of my existence since college, but I think
>I'm finally starting to grasp it.

Personally, I think much better in terms of groups of vectors than a 
matrix.  My math isn't nearly as advanced as many people think, given my 
perceived status as a "graphics god".


> > I just got a quote from KVH yesterday for their new 6DOF FOG IMU system --
> > $25,000 and 10-12 weeks delivery.
>
>Ouch. Would this be their "tactical grade" IMU?

Yes.


>I'm starting to wonder if our IMU will be OK for Gizmo, and I'm sure
>that it won't cut it for POGO and later vehicles.

We did a couple flights of our very first vehicle frame with tuning fork 
gyros before moving to FOGs, so you should be able to get some work done 
with the cheap xbow, which is a good step up from the gyration units we 
started with.  We did have to make a very vibration isolated container for 
the gyros, with soft foam pads supporting it on all sides, because any good 
thump would confuse the gyros.  The 10hz bandwidth was an issue for us when 
we were using solenoid based differential engines, but a jet vane vehicle 
should be able to fly fine with 10hz gyros.

I'm still willing to loan ERPS my old KVH gyros if you want.  They are 
discrete single axis analog output FOGs, which aren't nearly as convenient 
as the integrated IMU, but they do have 100hz bandwidth and are vibration / 
shock / G insensitive.  They were fairly expensive ($1500 each), but they 
are now a discontinued product.  You could have them on loan if someone is 
willing to take responsibility for $2000.

John Carmack