Yessir, as the title suggests – occupant-awareness of functional!

It works, surprisingly.  The gotcha now is that range is limited – it will happen multiple times overnight that the system will switch from Home to/from Away, as it has trouble finding our phones.  But I knew that that might happen – part 2 of this project is hardware in nature, to modify my USB Bluetooth dongle with an external antenna (and hence, greater range).

But ya – it’s uber-cool.  As always there were/are some bugs to work out.  Not with the detection steps per-se, as that seems to be rock solid.  Rather, with the tie-in of the surveillance system.  I had to think for a while about what and when a state change should occur.

Here’s the thing.  While it may be sufficient to poll to see when occuapants have left, it’s not really sufficient to poll to see when occupants have arrived.  By the time that a polling run occurs, the system may have already detected me or my wife coming home and send an alert.  And it may continue to send alerts until the polling run determines that our phones are home and – hence – the system state should change to Home.

That’s one.  Two, is that it’s very possible for either myself or the wife to forget our phones at home when we leave (you can determine who’s more likely to do this…)  It’s possible to set the system to an explicit Away state during which it will not poll for Bluetooth radios… but the problem is that the system (as currently designed) can’t expose this state to the polling process.

That means that the polling process is only used to determine when we’re away.  To determine when we’re home, the system relies on a surveillance event.

Which is fine – in theory that’s a great solution.  In practice we’ve got some implementation issues.  One, is that the limited range of the USB Bluetooth dongle means that an audio event will likely be the one to switch the system to Home – and audio events are most likely to fire for sounds inside the house (this isn’t always the case, but is a guaranteed case).  So an alert will still be generated due to exterior motion detection.

That problem can be solved with an increased Bluetooth detection range.  So we’re good there.

BUT… the reliability of the cammies’ detection events must be high if we’re going to use them to set the Home status.  That’s what I’ve been struggling with.  Not that they’re not reliable – rather, it’s the age-old problem of some required process going into La-La-Land(tm) and messing everything up.

And let me tell you, I’ve struggled with those problems over and over and over.  I thought things were pretty stable, until I found out that audio monitoring likes to… stop… …. .  Anyhow I think I’ve got that one worked out now.

So ya, once I get that external antenna hacked together we should be good!