room occupancy sensing

kwilcox

Active Member
I need a way to sense room occupancy that doesn't depend on motion within the room itself. Pressure sensors are out too. What's the state of the art here I wonder?
 
There are two other ways to detect room occupancy. One method uses ultrasound to send out in one direction than measure the bounced sound from the other direction. In theory, that is the optimized "distance". The sensor must be able to detect long period of time of no change everyday, Then any change against that long period of quiet time will be considered "occupied". Because it calculates the quiet time each day, moving furniture will affect mostly one day. Another method uses capacitance change to detect the room occupancy. Both method assume human will move at least every 4-6 hours.
 
There was an article that described using a Xbox 360's Kinetic sensor to not only detect occupancy but also control with hand motions. Pretty advanced stuff but super cool.
 
Most buildings use a combination of IR/Microwave and sound in their occupancy sensors. That's probably the best solution the industry has right now. I would think IR cameras would be the best solution. That is what makes something like the Kinect so appealing but the cameras on the Kinect have a limited field of view it probably wouldn't cover a whole room, maybe 10x10 if your lucky. There is a new one coming out for Windows that has a different camera setup, not sure of the specs.
 
IR camera/software that identifies heat signatures as people seems like the way to do it. You'd think that you could combine these into a self-contained device that would close a relay to indicate "occupied" within the sensor field of view. My research indicated that TI produced the base technology to create such a device almost 2 years ago:

http://www.ti.com/so...cupancy_sensing

I've not been able to unearth anything that employs this technology however.
 
There is a thread about using beam break detectors in a door jamb to detect the direction of a person entering or leaving a room and setting its occupancy state it relies on two ir leds and two ir sensors and a microcontroller.
 
thanks, I've been looking for that thread!

edit: damn, no circuit info posted.

Reading the description on his web site and coupling that with the PCB photos it's fairly easy to work out what he's doing, and from what I can tell it's almost all software on the PIC16 he's using, including generating the 38Khz pulses for the transmitter LEDs. The large IC is obviously the PIC16-series microcontroller, I assume the 8-pin IC is a MAX485 or similar for interfacing to his RS485 network, the transistor and resistor are used for driving the IR LED's, there is a voltage regulator and accompanying filter capacitors for supplying power, and then just a couple of LEDs with current-limiting resistors for activity indicators.

Here is the link to the full write-up on his web site in case you missed it:

http://www.mysmarthomeblog.com/page--16.html

There is also a "Contact Me" link on that site with an e-mail address, maybe you could get him to share his code and/or schematics?
 
This got me interested enough to do some research.
I don't need direction indication or body counting, I need threshold entry. My laundry room has an occupancy sensor, but the cats trip it. Withsomething like an IR barrier at 40" I can turn the light on when breaking the beam and off when the occupancy sensor no longer detects movement. So that cats will never turn on the lights. I've got the same problem in the master closets, we never shut those doors and the cats go in there all the time.

I've used 555 timer circuits for IR remote receving and retransmission.
So I looked for some 555 circuits.
This looks like about the best one for my application.

http://electroschema...-light-barrier/

I can increase the size of the resistor on the IR LED to reduce the drive and current consumption since in only needs to go 3 feet.
I can and steal power from the 12VDC going to the occupancy sensor and run it through an LM7805 voltage regulator and a couple of smooting capacitors to get met the 5VDC I need.

http://freedatasheets.com/blog/uploaded_images/7805datasheet-730100.gif

I can build both the transmitter and receiver on the same board with a 556 for the transmitter and a 555 for the receiver.
The receiver 555 can be set up as a one shot with a long output pulse, like 1/2 sec to a second, and send that pulse to the coil of a relay instead of the buzzer to provide a dry contact output to an Omni Zone.

You could go simpler, but the pulsed IR will give good immunity to stray light. The inset receiver will also give good protection from stray light.
The whole thing should only be a few dollars in parts.
 
I'm also working with Lyrtech currently to get my hands on several IOS eval units. These are non-motion dependent occupancy sensors that work by identifying human shapes based on their IR signature.
 
I'm also working with Lyrtech currently to get my hands on several IOS eval units. These are non-motion dependent occupancy sensors that work by identifying human shapes based on their IR signature.

Guessing those will run a bit more than $20 each, though :)
 
Back
Top