RE: Occupancy Sensors and No Light Switches

[Nigel Orr <Nigel.Orr@xxxxxxx>]

At 17:35 22/06/00 +0100, you wrote:
>levels, whether you've just come in from outside etc. etc. A discreet
button
>somewhere to say "It is now dark" and "It is now
light" is perhaps a good
>idea.

It's surprising how 'relative' sense of light and dark are- it's a bit like
heating, and is one of the things which makes 'proper' automation trickier
than it looks... perception of sound seems to be quite well researched and
understood, so it's easy to model, but heat and light seem to be tricky!

>enough. I found that standard alarm sensors tend not to be - they are
>usually positioned to be most sensitive to intruders coming through
windows
>etc, not people walking through internal doors.

The usual positioning is 'not pointing at a window' and 'set so that a
person
entering by a door or window will be walking across the sensor's view'.  In
my case, that almost always positions them so that when you open a door
into the room, it blocks the sensor until you walk further into the room,
so it would have too much of a delay to be usable for automatic light
switching.

The only way I can think of to solve the problem is to have a separate HA
sensor, positioned so it gets someone entering by the internal door
immediately.  But that means more 'zones', though the sensor wouldn't have
to be as high quality as standard alarm sensors, as the penalty for 'false
alarms' is much less.

>Knowing when to switch off
>lights is much harder.

My best guess for 'room empty' would be when the power consumption drops to
near zero... eg TVs, reading lamps etc go off.  If you think about it,
there aren't _many_ occasions (there are some) when you sit in a room with
no electrical appliances turned on.  But there's still a need for some sort
of manual override, and it does need you to have manual control over the
other appliances, which isn't very sophisticated!

Personally, I can't see myself removing lightswitches for 10-20 years, if
the house is ever to be sold, or even if it's just to be 'easy to use' for
visitors, they've got to stay- it's surprisingly disorientating if you
reach for a lightswitch and it isn't just where you'd expect it- especially
if you are in an unfamiliar place.  Any automatic system will have to be
over-ridden by manual lightswitches.

>  A person sitting still is invisible to most sensors I
>could think of.

I think Keith posted a while back about a more sophisticated one- basically
an IR-sensitive camera with some image processing to decide if 'heat blobs'
were human.  It should be something that can easily get cheaper, maybe in 5
years.

I know from image processing I've done that it's possible to make a _very_
sensitive movement detector, just using a standard CCD camera, without
_too_ much intelligence, but it does still need a PC per camera (it was
written in a lunchtime, so it could definitely get more efficient!)

Actually, on that topic, I was thinking about the 'detecting movement with
an outdoor camera' thing that was discussed a few weeks ago, it might work
reliably if you have a light sensor and _only_ look for movement between
frames if the overall light level doesn't change.  That should avoid
effects of the sun coming out from a cloud etc, and should be fairly easy
to implement.

> Another problem is that X10 isn't a reliable transmission protocol.
Signals
>sometimes go missing, which is also annoying 'cos the lights don't come
on.

So don't use X10- get some wires in! :-)

Nigel


------------------------------------------------------------------------
Click here for a chance to win airfare to Vegas for you and 20 friends,
$15,000 and a suite at Bellagio for New Year's from Expedia.com. Or
win 2 roundtrip tickets anywhere in the U.S. given away daily.
http://click.egroups.com/1/5296/9/_/2065/_/961746982/
------------------------------------------------------------------------