human detection

horizontech · Joined: 09 Nov 2003 Posts: 13

Hi !

I Looking to find a low cost solution to turn on LED on display when client are in front ( max 1 meter ).

I need to find some idea about ... IR.. Photocell.. or other

Thanks for your time

Alain Tanguay

libor · Joined: 14 Dec 2004 Posts: 288 Location: Hungary

There's been a short thread here about it already. here
There's a range of methods you can start to look for: PIR (passive infrared detectors), microwave alarm sensors, ultrasonic, IR laser remote thermometer, air movement (pressure) detectors, and don't forget a simple pressure sensitive door mat placed properly, or a light curtain with IR laser and mirrors. Or a much more resource demanding solution would be to process a digitized camera image...
The challenge for the alarm-type sensor solutions is to make enough hysteresis (in software I think), so you can detect a person in a temporarily moveless but still present state.

Ttelmah · Guest

A huge amount depends on the enviroment involved. Standard PIR modules are very cheap, and work quite well, but require 'movement' of the target. However these tend to miss-trigger in areas with bright sunlight, or rapid air movements (doors opening/closing etc.). Radio gets expensive, and runs into licensing issues. I'd suspect the easiest way would probably be ultrasonic. Small 8pin pic, generating a burst tone into a ultrasonic transmitter, receiver connected to a fairly simple active filter, and the output of this driving back into the PIC. System looks for the first return signal, after a blanking period designed to ignore any directly coupled signal inside the box, or through the PCB. I did a little rangefinder like this some time ago, and the whole thing involved a couple of op-amps, one inductor, and a 8pin PIC, and gave quite effective rangefinding over distances from about 6", out to typically 10 feet. The best 'test target', is a tennis ball, which is suprisingly difficult to detect (soft surface, and convex, so the return is very low). A human is comparitively easy.
You could do a similar thing, using a IR transmitter sending a tone, and an IR receiver looking for returned signals. However the problem is distinguishing what is a human. With this, the only 'signal', is a change in amplitude of the returned pulse. A flat wall 10 feet away, might well give more return than a human a couple of feet away. Without any form of 'distance' indication, workin out when a person is present, then becomes hard...

Best Wishes

jahan · Joined: 04 Apr 2005 Posts: 63

I've tested the PING)))) from parallax and it works pertty well for sensing movement (change of distance from previous ping). It's very easy to use BS2. I've tested it with 16F84A and gave pertty good result.

Price is not bad.

http://www.parallax.com/detail.asp?product_id=28015

valemike · Guest

jahan · Joined: 04 Apr 2005 Posts: 63

there is a ticking sound, but the room has to be VERY QUIT to hear it.
in normal operation, the blinking LED is more noticable than the very soft clicking sound.

newguy · Joined: 24 Jun 2004 Posts: 1909

Sharp has a number of analog and digital infrared distance measurement sensors. I've used them pretty extensively, and they work very well. They're borderline cheap (~$20 CDN), but they can measure distances out to 1.5 m (about 4.5 feet). Might work for you.

Go to http://www.digikey.com and search for 425-2062-ND
(Sharp distance measurement sensor GP2Y0A02YK0F).