Okay, okay.

I'm back to my original idea of monitoring the signal difference from the
transmitting wire on (or in) one wall to the receiver wire on the other.
Something like they do for traffic lights.  Signal changes when a body
occupies the room.

Seems to be even more practical now if you start setting up all this IR
detectors all over the place.

John Mullan


-----Original Message-----
From: pic microcontroller discussion list
[mailto:PICLIST@MITVMA.MIT.EDU]On Behalf Of Ken Webster
Sent: Friday, February 18, 2000 11:30 PM
To: PICLIST@MITVMA.MIT.EDU
Subject: Re: SV: Re: Home Automation with PICs [OT] (was: Internet
Toaster [OT])


>ok, but what if the room has 2 doors ?

The general case would be to assume that the subject(s) is/are in the
vicinity of the most recently tripped sensor or in any area lacking coverage
that lies beyond that sensor.

Thus if
   R = sensor in room
   D1 = sensor near (just outside) door #1
   D2 = sensor near (just outside) door #2

Then the following conclusions may be drawn from the following sequences of
sensor interactions:

D1 first then R --> subject entered room via door #1
D2 first then R --> subject entered room via door #2
R first then D1 --> subject exited room via door #1
R first then D2 --> subject exited room via door #2

Of course, a system this simple could be fooled by someone else simply
walking past the door when someone is alredy in the room and sitting still.
If you used more sensors you could track each individual and conclude that
it was merely another individual walking past the door.  Of course, this too
could be fooled if two individuals entered the room together, one sat down,
and the other left.  But at least it is a starting point.

If you wanted a perfect system you may have to make people wear PIC-based ID
transcievers that identify each individual to each sensor they pass (via a
radio signal).  You could even implant the chips under the skin like they do
to track wildlife.  Maybe the government will like this idea too and require
that each individual be implanted with such a chip at birth so that thier
location can be unambiguously resolved at all times :o)

A less invasive but more expensive approach would be to place cameras at
various locations around the house and have a computer vision system pick
out the human subjects and track them from room to room.  I don't think it
would be practical to attempt this using a PIC though :o(  ... maybe if the
human subjects had a big bar-code tatoo on their foreheads :o)

Cheers,

Ken