absolutely!  one of the easy (it can be done all with analog circuits or
any micro) ways to track an object that puts out "x" signal is to have 2
detectors, one on each side of a barrier (i.e. a photo detector on each
side of a piece of cardboard or projection from the surface) and steer
the mechanism so that the signal stays at the same level on both
detectors.  if the object moves left the barrier cast a shadow on the
right sensor.  to make the signals equal again you turn left and vice
versa.  i've seen this done purely analog to track the sun as it moved
through the sky for an electronic sundial that rotated on a platform as
the sun moved with a pointer to the "sun" time.  to track in 2
dimensions you use 4 detectors in the corners of an "X" shaped
projection or a 4 segment detector is sometimes used to track a laser
beam from the target, keeping the signal equal in all 4 quadrants.

of course most targets aren't kind enough to aim a laser at you so you
can track them, but you get the idea.  a more advanced system can use a
camera and just try to keep the object/light/dark source in the middle
of the image, but really just using 2 or 4 detectors is often enough and
takes very little processing power.  of course with a camera, or even a
couple of sensors you can look for something that doesn't move with the
background as you change direction to find the moving object in the
field of view.

with one detector you can put a tube  or circular block in front of it
and wobble back and forth to see what direction the target is at.  there
are a lot of such games that can be played.

if you want to be really low tech, you could follow the governments
example.  a project was recently declassified where they tried to use
trained pigeons to guide missiles, the birds were in the missiles
looking at a screen and would peck in the direction the missile needed
to turn (this went of for years, close to a decade i think).  they got
some pretty good results but eventually electronics started doing
better.  no joke, they actually had birds in missiles steering them
after training the birds with a monitor and reward system to peck in the
right direction to keep an image centered.  sorry i don't have the
details handy, but i bet a google search would turn up something.  the
project was declassified a year or two ago and there had been some
suspicion that such a program had existed prior to the release (i think
the article i read was in one of the electronics industry magazines).

i've considered making a robotic toy for my cat for years.  it would
make an excellent project (i think there is a cheap cat toy that does
this, though i don't know if it has any sensors or just moves randomly).
 basically an electronic mouse, preferably with an internal mass on a
motor so that when caught it could try and roll over or shake loose,
though of course you'd want to make sure to limit the forces generated
so as not to hurt a cat that was really getting into it....

Lindy Mayfield wrote:
------.
>
> Does this mean that my PIR motion sensor on my little robot can be used not just to detect motion, but as well to find warm things?  So I could put it on a servo and it could hunt down my cat?
>
> I have some experimenting to do!  Thanks!
-----------

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