In Bed Monitoring

[mrhappy]

Guys

Does anyone have any ingenious methods to monitor whether a bed is occupied or not? I know historically pressure pads have been used but they are not suitable for my particular setup, too unreliable and give me frequent on/offs.

I've probably been thinking about this little problem for a good while, I have thought about

1) some sort of distance temperature monitoring, the bed should have an ambient cold temp if no one in it - something like http://www.adafruit.com/products/1747 but not sure that the distance would be suitable. My nearest point it could be mounted would be about 5-6' away.

2) ultrasonic sensor above or the bed perhaps even adapting one of the car rear parking sensor kits, relatively narrow field of view and don't want to have to cut multiple holes in the ceiling really. Issue would be if I left something on the bed then this may give a false trigger.

3) strain gauge under the legs - perhaps even on all four legs (wheatstone bridge into an instrumentation amp and ADC) but the legs are onto carpet so it is unlikely that I could get a reliable indication.

4) I don't know what the technical term is but the bed has slats down the middle of it like the attached picture (that's not my bed, just a picture from Google). I figured that when I, a frequenter of bakeries, lie on it then the slats should depress to some degree and this should be noticeable to the point it could be sensed.

As luck would have it I had a triple axis magnetometer in a box of on-a- whim-ebay-purchases. Now this might look like a compass attached to a ruler I found in box of stuff of mine from school and that is indeed what it is. I have screwed it across three slats and then into a position of where I thought there would be the most depression .

I just put the angle values into a web page running on an Arduino and bring that into HS. I've been plotting the values and whilst it is detectable when I get in I obviously move about and then the sensor gives normal no-load values during the course of any sleep pattern.

Now I am back to square one, I know that you could probably deduce that people are in bed by other means (no motion activity, a manual button press etc) but that is no fun.

Anyone any suggestions?

[Rupp]

The elderly care homes use pressure monitors for this and it works really well if you have the correct mat.
https://www.youtube.com/watch?v=5H-KWeD251k

[S-F]

I have been using a pressure sensitive mat with a de bounce circuit I found here for years with great success. It works perfectly. That said, I'm going to scrap it when the Luna mattress cover hits the market.

[Mattyjee]

I have a motion sensor in the bedroom (under the bed pointing out) and one in the hallway. If the one in the hallway triggers first, then the one in the bedroom, then it proves someone is in the bedroom. If the bedroom motion goes quiet, the bedroom door (to outside) has not opened and the hallway motion has not re-triggered then it decides that someone must be in bed (or the shower or toilet) but it is enough for what I wanted it to do. I don't have any lights in my bedroom/ensuite hooked up but if I did I would use this as proof of bed occupancy vs shower/toilet.

My failsafe method is that I have the blinds in the bedroom automated and I have them open each day and the wife can't sleep with them open so when she presses the button at the end of the bed to close the blinds then the house knows she is in bed. The key to installing a button that you want someone else to press is to hook it up to some other function that makes them need to press it.

[LeoS]

How about a couple of CO2 sensors near the headrest; along with the load/angle sensors?

I think the industry standard is to use those weight sensor strips... But they need to be replaced every x months.

[mwaite]

I am anxiously waiting for the Luna Mattress. I believe there will be an API for it. Another crowd funder for me .

https://www.indiegogo.com/projects/l...martbed#/story

[mrhappy]

Cheers for the replies guys, I have a few pressure mats around and I guess a debounce circuit might work. I don't want the mat on the top of the bed so it would need to go underneath, not sure whether the weight of the empty mattress will still trigger it though. CO2 sensors are a good one though, not thought of that and guess that there would be a rise in the room, whether this is detectable though is something I could try.

I should've mentioned I knew about Luna (and would have no hesitation writing a plugin for it) but I think it is a bit expensive for me, won't fit my bed and whilst I don't usually give two cares for safety markings I don't see any applicable to the EU and am a bit mindful over heating+fabric.

[S-F]

CO2 monitoring won't work if the windows are open or you're doing a lot of cooking or any other number of caes. Doesn't sound too reliable to me.

I have a pressure mat under my futon so I imagine that the weight of a mattress wouldn't matter much.

[mrhappy]

CO2 monitoring won't work if the windows are open or you're doing a lot of cooking or any other number of caes. Doesn't sound too reliable to me.

I have a pressure mat under my futon so I imagine that the weight of a mattress wouldn't matter much.

I think the day I get so lazy that I am trying to do cooking in the bedroom is a sign I probably need to take a long look at life!

I'll give the mat I have a go again...

[jon00]

I tried a mat many years ago. I'm sure it had a fault. When I plotted occupancy, I sometimes got a graph similar to below

[sooty]

Depending on the construction of the bed and of course the budget, you could perhaps build compression type load cells into the bottom of the legs and with the help of an Arduino or something similar weigh the whole bed.

Paul..

[AshaiRey]

[LeoS]

CO2 sensor seems to work quite well when it's close (thus the headrest remark).

I have a NetAtmo in my bedroom (approx 20' x 25'), placed about 15' away from the bed (close to the TV and couch) and from the graph I can tell when someone is in the room (>2ppm), when the room is empty (0ppm) and when someone is probably sitting at the couch (>30ppm).

Granted I keep my windows closed and I don't cook in the bedroom.

Hm... I've yet to link this to occupancy flag of the bedroom. I should try it...

[mrhappy]

Excellent thanks for the ideas, I now have a legitimate excuse to buy some more parts although I think I have some piezos and sliding pots somewhere I can try with. At least if I had multiple flex sensors then I could use up multiple analogue inputs whereas starting to use multiple I2C devices is no fun.

[twa8000]

I have been using pressure mats hooked to a DSA-10. However now I have a Withings Aura. This system comes with a bed mat and is very accurate. It can detect when your in bed, when your awake and when your asleep. It also detects what sleep cycle you are in. They have an API and would be great to have this data in Homeseer, you can run events when you lay down or even when you fall asleep and wake up!

When I get some free time I might work on a plugin, however IFTTT supports withings but not the Aura, if they added support for the Aura we could easily get the data into Homeseer.