[sdiy] proximity detectors anyone ?

[harry bissell]

OK I have a lot of experience with ultrasonic detectors...

The most common frequency used is 25 or 40KHz because there is a
window at those frequencies where attenuation in the air due to water
vapor is a minimum. So, most of the detectors you will find will be in
those two frequencies.

You can do a 'time of flight' to get the distance.  You 'ping' the 
transmitters
in the glove, and wait for the pulse to reach the many detectors. Count the
time difference between the first and the second, and you will be able 
to tell
position.

It WILL be tough. The transmitters are narrow beam, with side lobes... 
you'll
need several facing all directions per 'glove'.  If you alternate the 
pings...you could use
two gloves.

You need to allow the 'ping' to die down. Ceramic transducers will ring 
for a many cycles after
getting an impulse.  The 'electret' types (Polaroid, and another company 
I could probably
find) have much less ringing. So the number of pulses is going to be 
limited to that... I'd guess that
you might be able toget a data point every 10ms or so.

The system will probably have to be pretty BIG... your movements will be 
in inches or feet..

If you are looking for smaller motions and distances, I'd suggest a 
Theremin (RF) or an IR
approach

H^) harry


jbv wrote:
> Hi list,
>
> today an idea arose during a conversation with friends...
> It began with proximity detectors used in robotics, and
> especially the fact that those sensors are used on a moving
> device to detect non-moving objects of the environment in
> order to avoid them...
> but what if the idea was reversed : fixed detectors to detect
> distance of moving objects ? for instance 6 detectors, each
> one on 1 side of a cube, and moving objects in between
> (a moving hand, balloons, bubbles, a butterfly, flies,
> smoke, whatever). Signals sent by detectors would be
> hooked to a uC to compute in realtime the xyz position of
> the moving hand, and these parameters would be used as
> VC signals for synths...
> in the case of a hand, I guess a joystick or a dataglove could
> also be used to determine xyz position in space, but I'm sure
> there are many other funny applications...
>
> so here's my question : does anyone have any experience with
> proximity detectors, and what would be the best choice : IR,
> ultrasonic...
> I've already found several papers on the subject via google, but
> I'd like to have the opnion of this list as well...
>
> Best,
> JB
>
>
>
>