Occupancy Sensing

[TonyNo]

This is an interesting topic.

I've used both Leviton ODC20-UOW and ODC20-UOW

Were there two different numbers at some point?

 

[123]

"In residential settings, these will pick up the motion of you turning the page on a small paperback book, typing on a keyboard or even just moving your thumb around to channel surf."

I rarely use this word but that is awesome!

Sounds like a fly could trip this thing. OK, maybe a horsefly.

 

[pkoslow]

Sorry,

It's been awhile since I'd worked with these, so found the models I think I used with a quick web search. I think the second model was a OSC10-U0W... looks like I originally pasted the same model twice.

The have a simple relay output so it's easy to connect to an input on Elk. The units themselves have some timer functionality that you can set so the Elk rules remain very simple (no need for timers, etc). When contact is closed, occupied. When contact open, unoccupied.

The devices ran on 24v DC so will require a different power supply. It looks like these might be obsolete models now? Be careful when you start looking at the Leviton occupancy sensors because some have line-level (high voltage) outputs to energize relays or fixtures directly.

It's my experience that a device like these are required to get reliable occupancy sensing and that no amount of mickey mousing around with PIRs and automations rules will provide acceptable results.

Cheers,
Paul

 

[pkoslow]

"In residential settings, these will pick up the motion of you turning the page on a small paperback book, typing on a keyboard or even just moving your thumb around to channel surf."

I rarely use this word but that is awesome!

Sounds like a fly could trip this thing. OK, maybe a horsefly.

Yeah,

They really are that sensitive. They're usually used in commercial environments and if you look around in office buildings you'll likely see them. They'll keep the lights on for that poor sucker still typing away in his cube at 3am. They're used a lot in restrooms too as they provide a very quick "instant on" when you walk in the door and will keep the lights on even if someone spends a very long time in a closed stall (they can hear what you're doing in there even if they can't see you).

For residential use there are some considerations. Absolutely no pet immunity is one of the big ones. A small dog or cat wagging it's tail will set it off. Usually in an office environment things are still and quiet when no one is there, and that's not always the case in a home. Creating lighting events with rules based on occupancy along with home/away, time of day, etc. will usually provide acceptable results.

Just like PIR based motion sensors, placement of the sensor comes in to play. These ultrasonic sensors can even hear around corners so you'll have to play around with placement to make sure the family room lights don't turn on every time someone walks down the hallway.

Cheers,
Paul

 

[drozwood90]

Does anyone have experience with or knowledge of the use of occupancy sensors for this type of application rather than motion detectors?

Please define "occupancy sensor". IMHO, there is no such thing as a single occupancy sensor as determining occupancy is way too complicated. The only real bulletproof solution is RFID tagging with many antennas and complicated triangulations algorithms. A rule based system can use a multitude of sensors (PIR, microwave, photo-beam, acoustic level, light level, pressure mats, pressure switches, capacitive loops) to try and cover most scenarios. For example, one could use a pressure sensor on a bed to determine occupancy. One could put a magnetic contact on the footrest of a recliner to partially determine occupancy. One could put a mercury switch on a recliner or rocking chair to determine occupancy. Some of these sensors are absolute on/off, some are only triggers for timers that eventually expire and the status becomes indeterminate.

They exist, they are just expensive.

I used to design premium thermal imaging cameras (real thermal, not IR).

There is software that can determine just about anything you want, however, you are talking 1K-2K per license (generally).

The issue is you need to have a staring sensor. It needs to understand what IS an empty room. And it needs to be adaptive for the scene (leaves, sun, wind blowing things around), to reduce or try to eliminate false triggers. I saw a package once that could detect people loitering, as well as leaving things behind (suitcases, etc.). It seemed to work very well, however needed tons of processing power.

There are cameras that can do this work directly and just trigger based on configured parameters (visible at 5K range, thermal in the 15K range), or server based solutions, which is probably 1-2K then 500+ per camera that you want to feed into it. This is the REALLY REALLY high end of what Collin sells.

I know for my house, I just use some cheap X10 wired cameras feeding into some cheap software that can do a "rough" compare of the room as well as watch for motion. I talk about the software on another thread here. I think it was called SuperCam (I don't have my server up yet as I just moved). If you are interested, I'll try to get the link to the software when I can get the name of it. The issue is, I only have one room like that, as I can't convince the wife to let me put more cameras up in the house. She does not like to be watched. So, outside, in garage, however, NOT in the house itself.

Oh well...that's a battle I don't want to fight.

I think the trick is to combine multiple sensor's inputs. Then you can get an idea about the room. For instance, I never though to put a door sensor on the flippy part of the recliner. Genious! However, I did make a pressure sensor that I put under the cushions on the couch. So, I can tell if someone is sitting there. I just couldn't figure out how to get a pressure sensor under the ATTACHED cushion of my recliner. So, having the sensor on the footrest...PURE genius!!

--Dan

 

[felixrosbergen]

How expensive are the sensors that pkoslov is referring to here?

The pet immunity is a problem indeed...although my cats are awfully good at laying competely motionless for hours on end....

 

[BrettS]

For bathrooms, I also have a door sensor that causes HomeSeer to keep the light on if the door is shut (with a 10 minute no-motion timeout to turn the light off in case someone just shuts the door). All of this works really well for us.

This is absolutely brilliant... I've been struggling with my bathroom motion sensors lately... because occasionally people tend to spend a bit of time in the bathroom and don't move much. Obviously the wife and guests do not appriciate it if the bathroom light goes off while they're in there, so I wound up setting the bathroom motion sensor timeout to like 16 minutes and it's still not a great solution... I've found that people still occasionally spend more time in there and the lights turn out on them, plus I hate leaving the lights on for 16 minutes every time someone walks into the bathroom to wash their hands and walks out right away.

And the best part is that my hawkeye motion sensor is right next to the door... and there's a nice hack that replaces the light sensor in the hawkeye with a contact closure sensor... I can put a magnetic contact sensor in the door and run the wire three inches to the hawkeye and it will not only notify me of motion, but also whether the door is open or not. Door open, the lights time out in 2 minutes... door closed and they can stay on indefinitely... or maybe 20 or 25 minutes just in case the door is left closed.

Did I mention that this is brilliant? Why didn't I see this post two months ago?

Thanks much,
Brett

 

[GraysonPeddie]

I'm just wondering...

Assuming you have a home automation controller that supports conditions/variables, can't you do something like this:

Assume the variable: bthrOccu

1. When a door opens and the bthrOccu is 0, set bthrOccu to 1.
2. If a motion sensor is tripped while the door is still opened, set the bthrOccu to 2.
This will ensure that the light will stay on.
3. Repeat steps 1 and 2 if the bathroom door is opened and closed (that is, if someone want to use the restroom).
4. When a door opens and the bthrOccu is 0, set bthrOccu to 1.
5. If a motion sensor has not been tripped for a while, set bthrOccu to 0.

Could this work?

 

[BrettS]

I'm just wondering...

Assuming you have a home automation controller that supports conditions/variables, can't you do something like this:

Assume the variable: bthrOccu

1. When a door opens and the bthrOccu is 0, set bthrOccu to 1.
2. If a motion sensor is tripped while the door is still opened, set the bthrOccu to 2.
This will ensure that the light will stay on.
3. Repeat steps 1 and 2 if the bathroom door is opened and closed (that is, if someone want to use the restroom).
4. When a door opens and the bthrOccu is 0, set bthrOccu to 1.
5. If a motion sensor has not been tripped for a while, set bthrOccu to 0.

Could this work?

That might work in some cases, but at my house the bathroom door is usually open unless someone is actually using it. Instead the first indication that someone was entering the bathroom would be the the motion sensor detecting someone. At that point I would turn the light on and keep it on for as long as motion was detected. If no motion has been detected for two minutes, then turn off the light (note that the door was never closed, so it's likely someone just went in the bathroom to quickly wash their hands or something).

In another case, the motion sensor detects someone entering the bathroom and I turn the light on. This time the door is closed, so now I won't turn the light off for 25 minutes after no motion is detected. That way if they're um... sitting still... then the light won't go out on them. Once they finish in the bathroom they'll open the door again and leave, leaving the door open. Since the door is open now the light will go out after two minutes with no motion.

On the rare occasion that someone closes the bathroom door when they leave the light will still go out 25 minutes later so it's still not on for too long. Actually, as I think about it, I could incorporate some further logic to fix this issue. If the door is closed and no motion detected after the door is closed at all, then the bathroom is likely unoccupied with the door closed. At that point I can probably shut off the light after two minutes of no motion. I'll have to try to test it a few times but given the location of the motion sensor I think it would be pretty difficult to not set it off at least once after shutting the door if you were in fact in the bathroom.

I guess your logic would need to be different (and probably closer to what you described above) if your bathroom door is normally closed when it's unoccupied.

Brett