[Message Prev][Message
Next][Thread Prev][Thread Next][Message
Index][Thread Index]
RE: Room occupancy detection and door opening/closing - 1-wire or
wireless solution?
>> but it can't tell you in what direction they're going, so
>> how will you know whether to count someone out or count someone
in?
You could get an idea of direction if combined with the PIR data i.e
If door state changed from closed to open AND the PIR was triggered within
the Room then you could assume someone was leaving the room. If on the
other hand the Room PIR detected nothing but the hallway PIR had been
triggered then assume they are entering the room.
Of course the accuracy degrades when you have more than one person in the
house..
>Beam-break sensors in the door jamb would signal someone actually
>passing through the door rather than just opening it. Multiple
>beam-breaks positioned horizontally could even give you the direction
of
>travel
Rather than beam-break sensors (which might be hard to fit i.e. having to
fit into either side of the door frames etc) I've toyed with the idea of
using Curtain PIRs mounted above on either side of the doorway, working in
a similar way but being possibly easier to fit. The one disadvantage might
be that it would "see" pets walking through the door, whereas a
beam-break could be set at a height higher than the pet.
Another thing I'm currently about to fit are some stair pressure mats under
the chair and settee cushions i.e. I can then detect if someone is sat on a
chair. I've done a quick and dirty trail run as proof of concept and it
seems to work OK, but I've no idea if they will stand up to the constant
pressure of someone sat on them.
Marcus
________________________________
From: ukha_d@xxxxxxx [mailto:ukha_d@xxxxxxx] On Behalf Of Paul
Gordon
Sent: 15 October 2008 10:17
To: UKHA Group
Subject: RE: [ukha_d] Room occupancy detection and door opening/closing -
1-wire or wireless solution?
Dermot, it strikes me that magnetic door sensors don't actually help
much in determining room occupation do they?
They just tell you if a door is open or closed. Unless you have
automatic closers on the doors, either could be a valid state regardless
of the room occupancy - my lounge doors for example are routinely left
open when the room is occupied. Even with automatic closers that just
narrows down the logic a little bit, so that you can know that if a door
is open, someone must be holding it open, & thus presumably passing
through... - but it can't tell you in what direction they're going, so
how will you know whether to count someone out or count someone in?
If you have more than 1 person in the house you can't rely on PIR's
alone, as they can't discriminate between the movement of a single body
or multiple bodies. Even with a combination of magnetic door sensors and
PIR's, I can't imagine how you could logically deduce room occupancy
with any degree of reliability, and certainly not in real-time.
I suppose if you don't care *how many* people are in a room, just
whether anyone is in there or not, it can be simplified a little, but II
sill don't think it would be easy (possible?) to make near real-time
judgements...
I am assuming of course that the purpose of the exercise is to determine
room occupancy in order to do smart things with automation of lights &
appliances etc. rather than you having some other requirement for which
you just need to determine/count door openings....
In order to do that a bit better, some other technologies, either
instead of, or preferably in addition to those mentioned would probably
be beneficial...
Beam-break sensors in the door jamb would signal someone actually
passing through the door rather than just opening it. Multiple
beam-breaks positioned horizontally could even give you the direction of
travel (and I guess vertical ones could tell you how tall they were,
which might be useful if there are kids in the house perhaps?)
Some CCTV systems have some image processing smarts in the software that
can count objects & people in this way.. - I know Geovision has
features
like this for example, so the software can be used (and is designed) for
exactly this purpose... - to count people in & out of a room to
maintain
a record of its occupants...
Not wishing to pooh-pooh the idea, but I've been around on this long
enough to learn that occupancy detection is pretty much the single
hardest thing to get right...
HTH
Paul G.
> -----Original Message-----
> From: ukha_d@xxxxxxx<mailto:ukha_d%40yahoogroups.com>
[mailto:ukha_d@xxxxxxx<mailto:ukha_d%40yahoogroups.com>]
On Behalf
Of
> dermot_bradley
> Sent: 15 October 2008 01:35
> To: UKHA Group
> Subject: [ukha_d] Room occupancy detection and door opening/closing -
1-
> wire or wireless solution?
>
> Hi folks
>
> I'm trying to figure out the easiest way to detect room occupancy and
> doors being opened/closed from a PC.
>
> I already have 4-core telephone cable round around various rooms and
> I'm in the process of soldering 1-wire temperature sensors to this so
> it would make sense to use this if possible.
>
> One easy way to wire up magnetic door sensors seems to be to use a EDS
> D2PC "low cost 2 channel digital I/O 1-wire card" (as sold
by
> Homechip) to connect the magnetic sensors to the 1-wire network.
>
> As for PIR sensors, I've not yet found a easy way to use them with
> 1-wire, mainly due to the requirement for power for the PIRs.
>
> I guess I could use a battery powered PIR and wire the contacts to a
> D2PC for 1-wire connection.
>
> Alternatively I could buy some Visonic wireless PIRs and their
> Powercode receiver to convert the PIR wireless signals to seperate dry
> contacts and then connect these to D2PC and on to the 1-wire network.
>
> Has anyone on the list tried this sort of thing before? Anyone have
> comments or alternative suggestions?
>
> Basically I'm intended to use the PIRs and door contacts so that a
> program on my homeserver PC can control lights and appliances (i.e.
> turn on room lights after X minutes of no movement).
>
> I haven't decided what to use for light and appliance control yet
> either (that's a whole other isseu) but will probably go with Homeeasy
> or Domia.
>
>
>
> ------------------------------------
>
> **** Sponsored By http://www.Berble.com ****
> **** Computers You Carry ****
>
>
>
>
>
[Non-text portions of this message have been removed]
------------------------------------
UKHA_D Main Index |
UKHA_D Thread Index |
UKHA_D Home |
Archives Home
|