Room level geofencing?

wkearney99 said:
For example, the front porch light is on a motion sensor and for a majority of the time that's a great feature.  But as warmer weather arrives it's nice to be out on the porch swing.  This, of course, would be a great time for an over-ride.  Trouble is, how do you structure this to get what you want, without unintended side-effects.  Or handle motion sensors for outside lights when you have a party.  A party is easy enough, if your rules are structured to take it into account.  As in, I've detected motion, but party mode is set, so nevermind.  This works OK for a party, but seems a bit tedious when you just want to sit on the porch for a while and not have the lights keep coming on.  What's the right "user interface" for this?  Using the wall switch for this runs dangerously close to the very annoying way old-school motion sensors would work; making you turn it on-off-on in some magic-dance-sequence to put them into a mode.  But the sensor is close enough to the doors that it triggers when they're opened.  Now, should it?  Or should it err on the side of restraint and not turn on the lights unless motion outside was detected first?  Ok, so instead of automatically getting lights when you go out, you have to press the button first.  But how long should it wait before resetting it's motion detection?  Once you're out there and the door is closed, how long should it wait before resuming a normal 'outside motion triggers lights' mode?  
 
I think this is the sort of stuff that can be done if you capture enough data.  Like you might not always sit out on the porch, but when you do it's normally between 70 and 75 degrees, after 8PM, not raining, and you're inside by 10PM.  Or is a Tuesday.  Or whatever.  Obviously it won't be perfect, but if you know your conditionals, then it's not too difficult to program them into your controller.
 
I don't think it would be 100%, but 80/85 wouldn't be that much of a stretch.
 
Rules are doable with one maybe two, but with 4 including a teen and preteen girl plus a dog it's crazy.

I have a decent vr system going already but I need to specify location, hence the post. And I struggle to remember what I called the room. (did I code MBR or bedroom or master bedroom or...)

If I could just say lights on, that would be moar better.
 
all of this is (as of yet) untested, but let's say that CastleOS tosses HTTP triggers as expected - you put a little $60 windows tablet with a kinect in each room, each tablet runs an instance of the CastleOS software to capture VR and send HTTP, you 'train' the CQC HTTP Trigger driver to accept input from a specific room-based URL, and then you tell CastleOS to send that HTTP trigger.
 
So if you're in the Living Room and you say Lights ON, CastleOS sends the HTTP Trigger specific to the LR over to CQC which would be configured to turn on the correct set of lights.
 
So, in theory, it would work exactly as you describe.
 
jkmonroe said:
all of this is (as of yet) untested, but let's say that CastleOS tosses HTTP triggers as expected - you put a little $60 windows tablet with a kinect in each room, each tablet runs an instance of the CastleOS software to capture VR and send HTTP, you 'train' the CQC HTTP Trigger driver to accept input from a specific room-based URL, and then you tell CastleOS to send that HTTP trigger.
 
So if you're in the Living Room and you say Lights ON, CastleOS sends the HTTP Trigger specific to the LR over to CQC which would be configured to turn on the correct set of lights.
 
So, in theory, it would work exactly as you describe.
Hoping not to buy single function devices like the kinect, or still yet more power drains like more tablets. If either the echo or Android with geofencing work, that would be ideal. At 1700 sqft but 9 rooms, clutter is evil. An echo is also a speaker so the wife might be able to handle that. (although in wall echo would really rock)
 
The more I see jkmonroe discuss the uses of http triggers, the more I wish CastleOS would be aimed as a supplement to any other HA systems rather than trying to sell itself as a competitor.  There is obviously a market for VR with HA and no one is selling a product that simply works with the existing HA solutions.  Especially one that's not cloud based (Echo or Siri)
 
That way the programmer works on what he knows best and improves it substantially (Kinect VR and possibly use the IR as well?) and lets other maintain the bazillion long list of drivers which ends up being the core of any HA software.  I mean with the IR and VR, its feasible the system could recognize an individual, taking IVB's scenario even a step further.  
 
I know he will not do it, but if he did he could develop it to a real high point.  I think he will be spending a lot of time reproducing what other systems already do instead of improving VR capabilities.
 
So if you're in the Living Room and you say Lights ON, CastleOS sends the HTTP Trigger specific to the LR over to CQC which would be configured to turn on the correct set of lights.

So now if we can get some company to makes decent MS units with smart microphones life could be much easier.
 
hmmmm...maybe a bulk buy on smart watches without the straps? :)
 
IVB said:
Hoping not to buy single function devices like the kinect, or still yet more power drains like more tablets. If either the echo or Android with geofencing work, that would be ideal. At 1700 sqft but 9 rooms, clutter is evil. An echo is also a speaker so the wife might be able to handle that. (although in wall echo would really rock)
 
The Echo is a cylinder about 4" wide and 10" tall.  It has a rubberized bottom pad under it to avoid interacting with surfaces below it.  Presumably for vibration, but it also does a good job of keeping it from being knocked over accidentally.
 
I am new here and cannot post links yet, however on instructables is a great college project using android phones to track cats. 
 
I have some old android phones that I plan to reuse a local control panels, with intercom and whole home audio.
 
You gave me the idea to go and add Bluetooth location tracking as well. 
 
I have some bluetooth tags that I will see if the "Ping" properly to use as trackers, but most people in my house keep their cell phones within arms reach. Also serve as a nice "Lost phone feature".
 
I can see BT working for room locations as it is short range but geofencing inside a home using GPS signals is not likely accurate rough to determine rooms with any accuracy.  GPS satellite signal errors are usually only good for 3-5m accuracy and not typically repeatable.
 
High accuracy take relative positioning GPS techniques that involves a corrected signal sending unit based on a known accurate position.
 
Then there is the signal reflections throwing off signal accuracy as is exemplified around large buildings and watching your GPS location jump from street to street and back.
 
IVB said:
Anyone figure out a way to have Android know what room of the house they're in (without using motion sensors)? Ideally in a way Tasker can detect. Bluetooth is too powerful. 4 smartphones and people in the house, so knowing per device preferred.

I could do an NFC tap but that requires remembering.
If bluetooth is "too powerful," is there an adjustment where you can turn down the Tx power?  I haven't looked at BT in depth, but all other radios I've looked at in depth have that capability.
 
One guy is using hiis fitbit and a couple BT transceivers to determine whether he's home or not.
https://www.youtube.com/watch?v=rqkT1Mrf_RY
So, if you don't want to reduce BT Tx power on your Android phone, maybe you could reduce it on your fitbit or siimilar
 
Back
Top