HA Rules - How do you handle manual overrides?

I actually spent years perfecting the bed sensor I use and I can tell you lots that DON'T work. I tried IR beams, stress sensors, even pressure mats and what may seem like a good idea in theory, doesn't work. The pressure mat is great to catch someone standing on it. Not so great catching a person horizontal on a bed. The only thing that worked for me, was replacing a bed leg with a device I built which basically measures the weight on the device. After it "broke in" its been 100% accurate for YEARS. Our house couldn't function without it.

Have any pictures? Or a how-to perhaps!
 
Lots of good ideas. Thanks.

Here's another solution that I thought of (tell me if this would work). If I use a multi-button scene controller in each bedroom then one button could be "Nap Mode". The HA software sees this event and stops controlling the room for a period of time...maybe for the rest of that day. Then the HA system resumes its automated control over the room. Another button could be to turn off "Nap Mode" just in case you want to resume automated control sooner.

The "take a nap" scenario was only one possible manual-override scenario. I suppose there would be others, too. Each case would have to be handled separately.

Pretty much what I posted... ;)


:) Yep. My bad for not crediting you earlier. I read your initial reply and thought "Hmmm...a multi-scene contoller? Good idea!". I think the pressure pad concept could be problematic, especially getting them positioned just right. There's a big difference between how I lay versus how my 10 year old son lays. An explicit "Do not disturb" mode makes sense to me.
 
... What if I'm sitting/laying on the bed watching tv? That's different than sleeping in terms of the type of interruptions you want. ..
Indeed it is and that's why it wouldn't have to guess because there are plenty of available parameters to help make a good decision:
  1. Bed occupied: Yes
  2. Lights On: No
  3. Motion: No
  4. Time of Day: Evening
  5. A/V Devices Active: Yes
The first four items lead one to to conclude that the person is asleep and the lights should not be automatically activated. However the last item suggests the person is lying in bed and watching TV/listening to music so if the lights could be allowed to turn on automatically.

Is it cheaper to hard-code everything to a button? Unquestionably. However, it is only one step above 'lights with timers'. There's nothing wrong with that (hey, I use it) but, as mentioned by others, if a home is to be truly "smart" it must, at the very least, react to, and optimally anticipate, a person's actions.

In the old Premise forum, someone had posted an excellent description of a "Zero User-Interface" for media components. If you insert a DVD in the player, you are implicitly indicating that you wish to watch a movie. The HA system anticipates your needs and ensures that all A/V devices are properly routed and the room's lighting levels are adjusted. It opened my eyes to the concept of a truly "smart" home that anticipates needs based on various cues and activities.

FWIW, the attached PDF contains the original post.

PS
I wonder if two pressure sensors, one under each of the bed's rear feet, would be an accurate way to determine bed occupancy? Or just one sensor and a shim under the other one to keep the bed fairly level. (EDIT: Woops, I wrote this before reading Monk's post about a custom bed leg.)
 

Attachments

I actually spent years perfecting the bed sensor I use and I can tell you lots that DON'T work. I tried IR beams, stress sensors, even pressure mats and what may seem like a good idea in theory, doesn't work. The pressure mat is great to catch someone standing on it. Not so great catching a person horizontal on a bed. The only thing that worked for me, was replacing a bed leg with a device I built which basically measures the weight on the device. After it "broke in" its been 100% accurate for YEARS. Our house couldn't function without it.

Have any pictures? Or a how-to perhaps!

Look on the HomeSeer board. I did post them there a few years ago. (Bed sensor, was a popular topic)
Basically it was made out of wood. A wood square with another wood square above it that rode up and down on four nails, one in each corner. Springs from Home Depot kept the top square of wood up, and the removed bed leg stub sat on this top piece. Then a switch above would sense when weight was removed. The key was that the switch would sense when weight was removed, NOT when weight was added, so put the switch above and not below the upper platform.
 
... What if I'm sitting/laying on the bed watching tv? That's different than sleeping in terms of the type of interruptions you want. ..
Indeed it is and that's why it wouldn't have to guess because there are plenty of available parameters to help make a good decision:
  1. Bed occupied: Yes
  2. Lights On: No
  3. Motion: No
  4. Time of Day: Evening
  5. A/V Devices Active: Yes
The first four items lead one to to conclude that the person is asleep and the lights should not be automatically activated. However the last item suggests the person is lying in bed and watching TV/listening to music so if the lights could be allowed to turn on automatically.

Is it cheaper to hard-code everything to a button? Unquestionably. However, it is only one step above 'lights with timers'. There's nothing wrong with that (hey, I use it) but, as mentioned by others, if a home is to be truly "smart" it must, at the very least, react to, and optimally anticipate, a person's actions.

In the old Premise forum, someone had posted an excellent description of a "Zero User-Interface" for media components. If you insert a DVD in the player, you are implicitly indicating that you wish to watch a movie. The HA system anticipates your needs and ensures that all A/V devices are properly routed and the room's lighting levels are adjusted. It opened my eyes to the concept of a truly "smart" home that anticipates needs based on various cues and activities.

FWIW, the attached PDF contains the original post.

PS
I wonder if two pressure sensors, one under each of the bed's rear feet, would be an accurate way to determine bed occupancy? Or just one sensor and a shim under the other one to keep the bed fairly level.

I agree with you almost entirely with your Zero-UI world, but I have to tell you, we are in the minority. I use CQC, and from their board, 95% of the talk is about the Interface Viewer this and Interface Viewer that. I tell people I don't even use the Interface Viewer, but still have a fairly large and complex system, and I can tell, their eyes must be glazing over. I do use text-to-speech, and have some LED display boards, so I'm really not Zero-UI, but I'm close. 99% of the people there that do automation, instead of turning on a light, they would rather scroll through screen after screen, find their "light" button, then touch their touchscreen to turn it on.

P.S. No pressure sensors under the bed feet won't work. Been there, tried that.
 
...I agree with you almost entirely with your Zero-UI world, but I have to tell you, we are in the minority. I use CQC, and from their board, 95% of the talk is about the Interface Viewer this and Interface Viewer that.
Yes, it is unfortunate; glitzy user-interfaces have become synonymous with "Home Automation". I think the UI has captured far too much HA mindshare.

For example, visitors to my home love the kitchen touchscreen and ask "Is this a smart home?" ... because, of course, a home cannot possibly be "smart" unless it has at least one touchscreen with a sexy UI. Naturally, it is even "smarter" if it has a wall-mounted touchscreen. :)

I like the family photos displayed by my touchscreen but would prefer to never visit it in order to control something ... I want my home to anticipate my needs. I'm far from it but that's my goal.

... pressure sensors under the bed feet won't work....
Rats! ... and thanks for saving me the time and effort.
 
My solution is a bit hard-core. I count bodies in and out of rooms using two motion sensors angled slightly apart. Outer-then-inner adds to the room count, and the opposite of that decrements the count. When >0, light turns on. When 0, light turns off.

How the hell did you get that working? I tried to do that years ago! The hardware setup was near impossible!

Details details details, please!
 
A dedicated "I'm asleep" button is certainly one way to solve the problem. However, it introduces a step that is not normally involved in daily living. For example, you don't normally post a sticky-note on the bedroom door to siganl to everyone that you are going to sleep. If the entire household can adapt to this arrangement then OK. Otherwise, I prefer the "zero user-interface" approach where the house senses your activitues and acts accordingly.

Hmmm - I guess I disagree a bit - generally, if I'm gonna go take a nap, I tell anyone else in the house I'm going to take a nap - the implication is that I don't want to be disturbed. The button approach is like telling your house not to bother you.

Zero user-interface is good where it's a pretty much a slam-dunk (open a closet door and a light goes on automatically, close it and the light goes off). But it gets much trickier when the house has to "guess" at your intention. What if I'm sitting/laying on the bed watching tv? That's different than sleeping in terms of the type of interruptions you want.

You also want to factor in implementation cost in terms of time, $$, technological knowledge needed, etc. A simple button toggle is really cheap from this perspective.

well, with the pad, it's zero interface, and if you use the Harbor Freight pad, it's $11 + some way to sense the contact closure.

As for how to determine if you are napping, as 123 pointed out, if the lights are OFF and you are in the bed...
Or if you have shades, if the lights are off, and the shades are closed, AND there is bed pressure, leave the lights off.

IF you had the lights on because you walked in, then went to the bed, sat down, leave the lights on...then when you are ready to sleep, turn them off...
 
...I agree with you almost entirely with your Zero-UI world, but I have to tell you, we are in the

.. pressure sensors under the bed feet won't work....

Most pads don't work...BUT if you put a cheap-o pad, like the Harbor Freight one, it's sensitive enough that you can put it between the box and the mattress. In order to make it more sensitive, put little mountains of melt glue on it, so it can get pressure points of "contact"

When I tried it, it worked pretty well (without the glue). Someone on the HS board told me about the glue, seemed to work for him. I have not had a chance to do this, as I'd like to get 3 more pads (two for the wife's side, and one more for mine).

--Dan
 
Low tech approach...walk in the room, put a piece of tape over the sensor, take a nap, wake up, take off tape.
 
OK, please bear with me, folks. I'm still struggling to understand how to handle manual overrides. The nap scenario is just one use case. So I decided to go back to the drawing board on my original question.

Let's say that you have a room with an occupancy sensor that turns the lights on when you enter the room. Now, you enter the room and the lights turn on (good so far). You manually turn off the lights. Then, as you move about the room the occupancy sensor would detect you and turn the lights on.

So to handle this, would you set some kind time-based flag to know that the HA software should not turn on lights if they were turned on within X minutes?

Now, let's apply this to the nap scenario... You walk into your bedroom and the lights turn on...you manually turn them off so that you can take a nap. However, by time you reach the bed the occupancy sensor has seen you so the lights turn on. (I am picturing an Abbott and Costello skit here)

So we would need to tell the HA software to ignore any motion for some period of time. Right?
 
Manual intervention (turning the light on or off by hand) signals the HA system to behave differently. Turn off the light by hand and you're saying "I really want this off now. Don't turn it back on!". The HA system will suspend automated light control.

The next question is "When should it restore automated light control?" and the answer depends on what you want. It can restore automated operation after a predetermined time period (not ideal) or by looking for behavioural cues that indicate you want it to be restored. Perhaps a combination of time period, time-of-day, activities inside and/or outside of the room ... or just a hard-button by the bed-side that indicates "OK, as you were!".

The logic for controlling a "smart" light can become quite complex.
 
Pressure pads seem like they should work, but if you see what's inside them its foam, with some circles cut out and two conductive sheets on both side. To trigger them you must press down hard vertically so the two conductive sheets contact each other. But this under a bed leg, and it would always be triggered. Even with shims, adjustments, its hard to get it exactly right. I never could.

Put the pressure pad between mattress and box spring, and it will only trigger when getting into or out of bed, but not when your on the bed. The goal of a mattress is to spread out your weight across its surface. This is the opposite of what the pad takes to trigger, a heavy weight over a small area. I've tried pads on top of the bed, pads put on boards, even pads between two pieces of plywood between mattresses. It would never work accurately for one or two people.
 
Back
Top