> coil

I think that you will find it very hard to get the earth field signal with
many motors and other items around. Besides I think that two coils should
be used, both in a vertical plane, so a comparator could be used to get
the 'middle'.

I have played with the real thing (earthfield rotary inductor) and of
course everything magnetic will induce a voltage into it, such as your
keys, the structure under the tabletop, armatures in the wall concrete,
tools, etc. Not good.

Maybe with Hall sensors and some dedicated analysis better results could
be had. An optical horizon sensor is usually better afaik, assuming open
air operation.

I find that many people are mistaken about what exactly 'horizontal' means
in the evryday sense. If you'd manage to make a robot keep 'horizontal' in
the geodesic sense and it would have to climb a 15% ramp it would likely
fall over. Humans use both geodesic 'vertical' (using the special organ in
the middle ear), and optical clues. When either of these systems fails,
serious troubles occur to the respective human. I'd take this as a strong
hint ;-).

Maybe all you need to determine vertical in a slow-accelerating vehicle is
a critically damped pendulum. The computer could take care to adjust
pendulum readings when accelerating or decelerating, and could learn when
it is on a slope by taking care to zero a (by stopping or by running
with constant v) and read the pendulum position.

Peter

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads