The Virtual Crib - Home Automation Software Forum Index
RegisterSearchFAQMemberlistUsergroupsLog in
Occupant Room Detection

 
Reply to topic    The Virtual Crib - Home Automation Software Forum Index » Feature Requests View previous topic
View next topic
Occupant Room Detection
Author Message
Talryyn



Joined: 23 Jan 2009
Posts: 34

Post Occupant Room Detection Reply with quote
I just thought of something, it was discussed in another thread but my thought is a bit different.

The thought is instead of using pressure pads or motion sensors to detect moving objects. Why not use two optical sensors in the door jam, or just outside the arc of the door.

So for example, you have two sensors

X X Door

And when the sensors detects

X --> X Door

It knows one person entered the room.

X <X> X Door
X --> X Door
X --> X Door
X <-- X Door

So we had three people enter, and one left - leaving two people still in the room.

You can place the sensors high enough to avoid cats and most dogs, unless you have a dane or something.


What I could see happening, is since we know someone entered the room and has not left, lets activate a different room temperature setting, so if they are sleeping it will keep the room at an "occupied" temp. If nobody is in the room, then we can let it cool off/heat up by a few degrees. Saving a percentage maybe on your utilities.

Just a few thought, getting ready to head out with the family, I will clean up my post when I get back, add some pictures, find the sensors etc.
Sat Jan 24, 2009 11:17 am View user's profile Send private message
hobbes487



Joined: 10 Oct 2007
Posts: 268
Location: Boston, MA

Post Reply with quote
Sounds like a pretty good idea. Are the sensors able to detect the direction of movement or just that something passed through the door? If not, there is no way to tell if someone is leaving the room or entering.

_________________
Brian
Sat Jan 24, 2009 1:08 pm View user's profile Send private message AIM Address Yahoo Messenger MSN Messenger
Humanzee



Joined: 22 Aug 2007
Posts: 519
Location: Seattle, WA

Post Reply with quote
That's why he said to use 2 senors. So you could tell which direction someone is traveling by which beam was broken first. This idea is entirely plausibly, however, wiring multiple optical sensors seems like a lot of work. I'd probably only do it for a few places, like the bedroom, so that the house knows that there are two occupants in the bedroom, therefore change mode to sleep. etc.

Actually, a conventional motion sensor can know which direction someone is traveling, however, they are not usually wired to make use of the that fact. When motion passes from right to left the output may be + then - polarity, from left to right it will be the opposite. Instead of just looking for a difference in the two elements, one theoretically could be wired to show direction or travel as well.

http://www.glolab.com/pirparts/infrared.html

The real sticker is with resolution and those situations where the sensors could become confused. If the doorway is wide enough for two people to enter side by side, or if a swinging arm trips a beam before your body, you may get a false count. Interesting though. If you go though with this type of detection, post your project. It will be interesting to follow.

vCrib should be able to handle what every you come up with.

_________________
vCrib Tester #1, Forum moderator, using INSTEON devices, X10 sensors and remotes via W800RF, All doors are wired with contact closures.
vCrib Wiki
Sat Jan 24, 2009 5:27 pm View user's profile Send private message
Talryyn



Joined: 23 Jan 2009
Posts: 34

Post Reply with quote
We are getting ready to buy our first house, so we are going to rip out the drywall and do spray foam insulation. So while the walls are bare I am trying to think of what wiring I am going to put up for future use, and I thought hmm what about optical sensors.

I thought about it at dinner and I have a few ideas, I will browse around and see if I can find what might work, if it can be used with fidgets, etc.

At my last job I got to play around with a video system that detected movement by looking at a change in pixels. Well I guess newer security video systems do this to only record when it detects something moving - at the time this was because HDD space was limited so to speak. Wow funny to think that was over five years ago already... The system we were using it in was also able to look at lots of zones, and in one case you could take a parking lot, see when a space was empty/full and then also track if that car had paid for parking. Not sure what ever came of the system, I guess you could use a webcam to do this as well.

However I think the best way might be to use two simple sensors, that way you can tell the direction of movement. Get it high enough so it only will see the body, but low enough for kids as well.

I love home automation, so I am looking forward to playing around with vCrib, and hopefully adding my input here to the community as well. Smile
Sat Jan 24, 2009 10:06 pm View user's profile Send private message
Humanzee



Joined: 22 Aug 2007
Posts: 519
Location: Seattle, WA

Post Reply with quote
find Snyperbobs thread about his home wiring, there is discussion there about the things to wire in, even if you don't expect to get to installing them any time soon. I am lucky enough to have a good crawlspace that I can access so I haven't had much problems retrofitting wires later.

_________________
vCrib Tester #1, Forum moderator, using INSTEON devices, X10 sensors and remotes via W800RF, All doors are wired with contact closures.
vCrib Wiki
Sun Jan 25, 2009 1:53 pm View user's profile Send private message
SnyperBob



Joined: 22 Aug 2007
Posts: 630
Location: Illinois

Post Reply with quote
If your walls are going to be open, just wire for anything and everything you think you might want. I'm still here 3 years later and haven't put up drywall because I keep coming up with more idea. Guh....

I'll be pulling the trigger in a couple months so I can get this crap done. I'm getting sick of it, lol.

_________________
Visit the vCrib Wiki!

Clever image to be inserted here soon.... Smile
Mon Jan 26, 2009 4:49 am View user's profile Send private message Visit poster's website AIM Address
Talryyn



Joined: 23 Jan 2009
Posts: 34

Post Reply with quote
These seem interesting:

http://www.glolab.com/dp-005/dp-005.html

http://www.futurlec.com/PIR_Sensors.shtml

http://www.softlyspokenmagicspells.com/halloween/light_beam.html

Some of this stuff is actually pretty expensive (on the links I am not posting).

My thought is to build either 1) a module for one side of the door that has an IR emitter and receiver, on the other side of the door you have the same module but with the sensors opposite. So the receiver will see only the beam from that side. The emitter would be powered at all times, almost like a Wii sensor bar. When you walk through the door you break one beam a few milliseconds before we break the other beam. So we know an object moved in one direction.

2) Similar to a garage door beam sensor (thru beam?), we again use an emitter in this case an LED, and a light sensor.

3) Lasers!!! Wink http://www.windmeadow.com/node/5

Here is a usb IR sensor: http://www.toradex.com/@api/deki/files/76/=Oak_Move_Datasheet_V0110.pdf http://www.toradex.com/En/Products/Oak_USB_Sensors

Well now I am just going to keep looking, and I thought I found the end of the internet....

Edit: Finding more links with good info: http://www.glacialwanderer.com/hobbyrobotics/?cat=5 Scan down to see Laser Sensor
Mon Jan 26, 2009 11:07 am View user's profile Send private message
Humanzee



Joined: 22 Aug 2007
Posts: 519
Location: Seattle, WA

Post Reply with quote
It's fun to think about all the possible things that could be done. If you will be severely limited later, go ahead and put extra wires where you think you may want sensors. Keep in mind though that it is easy to get carried away with the grand plan in your head.

When we were remodeling I pulled wire to every interior door and wired in contact closures, and home ran them back to the server room, etc. Now a year later, I only actually use a handful of all those wires I pulled under the house. The laundry room pocket door is almost never closed, so why do I need vCrib to know its status? If it were closed, what real difference would it make to my automation? Don't get me wrong, I'm glad that I pulled the wire, I just haven't found a use for it yet. Some of these real world scenarios you will not need to ponder until you are farther into your system. But you should take some time to think about what you need to do instead of what you could do. Then prioritize.

You can easily invest a lot of energy into figuring out exactly how you are going to do you beam sensor, but you may over look the fact that the space on either side of the door is already going to be monitored by motion sensors and you'll have to open the door anyway. So a motion senor on either side of the door along with the door sensor can tell you the direction of travel with out having to re-invent the wheel so to speak. Also, once you have established a direction of travel, what are you going to do with that data?

In our home, I have not found a real good reason to need the ability to track occupants around the house with 100% accuracy to the point of being able to display somewhere the actual number of occupants that are in each room. I can think of some scenario's where it could be used, but in reality the end commands for when to have a device on or off are usually the same regardless of the number of occupants. Rooms stay "OCCUPIED" until some defined increment in time after the last occupancy triggering event has occurred. Of course you can force this behavior to change, but do you need to? In everything I have done to date vCrib hasn't needed to know the instant that a room becomes vacant.

I'm not trying to discourage anyone from building out hardware and experimenting with different configurations, I'm just saying that the more complexity you add, the more possible modes of failure you introduce. With a high number of standard devices it can become difficult to maintain your logic systems anyway. Point being, that there is a lot of thought that goes into planning an automation system. I would prioritize and make sure you take care of the major things first.

Do you have enough power outlets to run your gadgets? Is their circuit sufficiently sized in your breaker box?
Do you need to put in extra lighting or switched legs to outlets so that you can automate lighting or appliances later? (Christmas Lights, Yard Fountain)
Do you have the correct number of conductors pulled for your ceiling fans?
Do you need / want network drops for every room,
Do you need / want network drops for security cameras?
Do you need / want structured wiring to future proof specific places?
RG6?
Do you need wires for all your doors and window sensors, or are you going to use wireless sensors. Are you going to be happy with battery operated devices?
Do you want wired or wireless motion sensors? Do you need extra sensors to cover your space correctly.
Speaker wiring/positioning?
If you move later, will anything you are doing negatively effect re-sale value? I.e. If you have to remove your automation to make a sale, will it leave deal breaker sized holes in the wall for those things devices you can't afford to abandon?

Once you've done all that planning for the things that need to go in the wall, and selected your lighting control technology, and priced everything out, you'll probably be really happy that vCrib is free. Then start to think about the things you want, but don't really need and put them in a separate list. In a remodel you end up making compromises and changing your plans. My background is in construction management, and I know that there is no such thing as a perfect set of construction plans or specifications. Things will change and stuff won't work out the way you thought it would. It will take longer and cost more than you think.

Case in point, I wanted a nice 200 gallon fish tank for our home, I designed this sweet cantilevered stand and worked out how all the equipment would be concealed and engineered the structure to make it all work. But it turns out that the space I wanted to put it in is simply too small and wouldn't work by a long shot with the furniture we ended up choosing for that room. Fun to design, but a waste of time considering how many other important things needed doing.

So go ahead and design your dream stuff, that's fun. Make allowances for actually building it later, add some extra conductors here and there, but I wouldn't buy any specific parts or pieces of equipment until you are ready to install it and have a good deal of certainty that it will work exactly as you have planned. I have boxes of automation bits that I bought years ago thinking that I would use them all right away but didn't end up needing. You may end up finding out that the extra work does not add value to your system or its capabilities. Spend your extra money on things you interact with. Spend extra time with the spouse, then after they go to bed, spend it on things you can program in vCrib.

I thought about installing custom IR emitters in my base boards to navigate my Roomba into different rooms around obstacles and such so that I could more thoroughly clean each room on different days. I described the plan to the wife, and she said, "Why don't you just buy a couple more Roomba's". Though her solution isn't as fun, it does make more sense when it comes to money, effort, and the time it would take to pull off my grand plan.

_________________
vCrib Tester #1, Forum moderator, using INSTEON devices, X10 sensors and remotes via W800RF, All doors are wired with contact closures.
vCrib Wiki
Mon Jan 26, 2009 2:50 pm View user's profile Send private message
SnyperBob



Joined: 22 Aug 2007
Posts: 630
Location: Illinois

Post Reply with quote
Very well put Humanzee. This is what I ran into when trying to design a way to automate my HVAC. It became a complicated problem for my specific setup, which would have cost me over a thousand dollars to make something reliable. I just couldn't justify all the time and money that I would have to put into something like that, just to save a few bucks a month on the heat/electrical bills.

Basically, the best advice I can give you is to determine what your core features are that you want for your home automation. If you want lighting control and security system, etc...

Once you can say that you implemented those features 100%, then I would move onto more advanced stuff. Usually it's not that hard to tackle advanced projects/ideas.....most of the time you're held up by funds. That's the case for home automation. I would love to have all Insteon (or other) switches in my house, but I just don't have $1500 to upgrade something that already works fine. I have a few x10 controlled switches that were cheap. Yeah they suck and don't work 100% reliable, but I still have that $1500 in my pocket for now.

if you're going to run wires to every door jamb just so you could see direction of motion, that would cost a lot of money in wires.

But, if you're really wanting to implement this, I wonder if a buncha phidgets motion detectors would work. You could put these right in the baseboards:


_________________
Visit the vCrib Wiki!

Clever image to be inserted here soon.... Smile
Tue Jan 27, 2009 1:02 am View user's profile Send private message Visit poster's website AIM Address
SnyperBob



Joined: 22 Aug 2007
Posts: 630
Location: Illinois

Post Reply with quote
I just checked out this link that you posted:
http://www.futurlec.com/PIR_Sensors.shtml

lol, those are probably the same sensors (basically) as the phidgets one, just a lot less $$. Nice.

I was actually thinking of trying to implement a laser system into my home security. I didn't want to have ugly sensors on my doors. I was thinking about having a laser that crossed the door, with tiny mirrors reflecting around the floor. I could cover two doors with this by bouncing the laser beam around two corners. Something like this. I thought a tiny mirror to bounce the laser beam would be more attractive that motion sensors and contact sensors all over everything.



I could hide the laser behind something, along with the PIR behind something. The mirrors would be tiny one inch squares along the baseboards. Most people wouldn't even notice them and I wouldn't have to rip off trim to put recessed sensors in the one door.

This is what got me interested in doing something like this:
http://www.instructables.com/id/SYTR52PFD80BUX9/

I was thinking about doing something like this outside for car detection on the driveway, or to detect someone walking thru the yard, etc... I haven't priced it out, but it's an idea. I need to see if I could mod an x10 motion detector to get this to work. I can't afford a security camera system yet, so I'm trying to come up with other ways to detect people snooping around (we've had a couple problems with this already)

I just don't understand how to wire the circuit.....this seems very confusing to me and I've done a buncha electrical engineering classes:


_________________
Visit the vCrib Wiki!

Clever image to be inserted here soon.... Smile
Tue Jan 27, 2009 1:19 am View user's profile Send private message Visit poster's website AIM Address
Humanzee



Joined: 22 Aug 2007
Posts: 519
Location: Seattle, WA

Post Reply with quote
Good idea in theory with your lasers Bob. But I think your house would have to be built of concrete for that to work. I know our house moves a little bit when we walk around. It wouldn't take much deflection at that distance for the beam to wander off your target. Your driveway idea may work better, but there you have to deal with frost heaving, temperature swings, vermin etc. I imagine this is why the door chimes and garage doors use IR instead of laser. You don't have to be as precise. I've made and played with laser projector stuff, small movements of the mirror compound to large ones with more distance to your target.

_________________
vCrib Tester #1, Forum moderator, using INSTEON devices, X10 sensors and remotes via W800RF, All doors are wired with contact closures.
vCrib Wiki
Tue Jan 27, 2009 12:13 pm View user's profile Send private message
Talryyn



Joined: 23 Jan 2009
Posts: 34

Post Reply with quote
Here is what I am thinking.

----------------------------------------------------------
[ ]
[ ]
[ photocell laser diode ]
[ ]
[ ]
-----------------------------------------------------------

On the other side of the door you have the same sensor but opposite, with the whole sensor board being maybe an inch long. With the components being SMT they should be aligned correctly, just make sure they are at the same height. Some of the laser diodes are visible as well, so this makes it even easier.

I could see using the phidget board to control this, we would just need to assign which photocell is the first one.

This could then be mounted in the door frame behind the wood, with just small holes being visible. If the aim is correct, the hole can be the size of the beam.. lol
Tue Jan 27, 2009 12:54 pm View user's profile Send private message
Humanzee



Joined: 22 Aug 2007
Posts: 519
Location: Seattle, WA

Post Reply with quote
Is this with an actual door or just in a doorway? The actual door will otherwise cause one of the sensors to trip first regardless of direction. You'd have to stuff both sensors on one side of the jamb. I'd think you'd want them spaced as far from each other as possible in case there is any timing resolution problems. You could even put one through the door stop trim. That's going to be tight to build as you probably only have an 1-3/4" ish in which to fit two sensors. The frames are also usually only 3/4" thick so you'd have to hog out some room in the stud framing to fit a one inch deep sensor pack. I had to do that for my magnetic door sensors in spots where the wiring was tight.Right now we are only marking down to the second I believe, so your conditions for first this event then the other may falter anyway. I might be wrong about the timing resolution for events logged though, regardless it should be fixable for optical gates if enough people have interest in using them.

In the attached drawing I made a MS paint sketch. In that I put the sensors apposing each other, with that configuration, you could use IR or something so that you don't have to worry as much about aim. With them opposing, each sensors should only be capable of picking up one emitter. Alternatively you could use two IR emitters of different frequency, and put the sensors and emitters in the same jamb. Then you'd only need some reflective tape or something on the other side of the doorway.



doorway sensors.JPG
 Description:
Diagram.
 Filesize:  6.87 KB
 Viewed:  18531 Time(s)

doorway sensors.JPG



_________________
vCrib Tester #1, Forum moderator, using INSTEON devices, X10 sensors and remotes via W800RF, All doors are wired with contact closures.
vCrib Wiki
Tue Jan 27, 2009 2:13 pm View user's profile Send private message
Talryyn



Joined: 23 Jan 2009
Posts: 34

Post Reply with quote
Yep your drawing is much better than my ASCII drawing, that is exactly what I am thinking.

I was worried with the IR sensor getting a false detection since an emitter is next to a detector, but if it is embedded in the wood (behind it, whatever) then that problem is gone.

A neat trick with IR, take your digital camera and look on the view screen and you will see your IR emitters. You can use your phone camera as well.
Tue Jan 27, 2009 3:44 pm View user's profile Send private message
Humanzee



Joined: 22 Aug 2007
Posts: 519
Location: Seattle, WA

Post Reply with quote
Talked with Vaughn about the timing limitations, turns out its probably not an issue because of the way those conditions are validated.
Quote:
MySQL only stores to the second, so I select by the second, then subsort by ID so things come in order, but technically several entries could be at the same time. Once MySQL goes to millisecond, so will I

You can tell direction. you just use a standard Event1 within 1 second of Event2 and vise versa

I have contemplated that same project for some time. It still has its flaws, like if you turn around in a door way, or people cross paths in a doorway, it will still lose track. that happens in my kitchen a lot, people coming and going a lot


The key thing here is that event 1 is WITHIN _ seconds of event 2. A fraction of a second is still within a second so it should work depending on the order events arrive. So sub second events, should make it in to the log in the correct order by ID regardless of how close they are.

_________________
vCrib Tester #1, Forum moderator, using INSTEON devices, X10 sensors and remotes via W800RF, All doors are wired with contact closures.
vCrib Wiki
Tue Jan 27, 2009 4:27 pm View user's profile Send private message
Display posts from previous:    
Reply to topic    The Virtual Crib - Home Automation Software Forum Index » Feature Requests All times are GMT - 6 Hours
Page 1 of 1

 
Jump to: 
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You can attach files in this forum
You can download files in this forum


Powered by phpBB © 2001, 2005 phpBB Group
Design by Freestyle XL / Flowers Online.