HA: Is there really minimal "new" in last 7 years?

[IVB]

pete, i totally agree on the "VR is not automation", which is why I put it into the "method of access". If you recall, even in 2006 when I was doing all those HA webinars, the first thing I said was that pretty little screens weren't HA, they were alternate methods of manual control.  To me, automation is having a rules-based engine where the house knows what you're going to do so you don't have to do it.
 
But 9 years after that first delivering that presentation, I still don't see us being that much closer. My new VRCOP is coming today (from automated outlet, hopefully this one works as the one from amazon never did), I might do some cutesy nit stuff like "trap the code sent to the zWave door lock and use it to disarm the security system" but not sure what else.
 
In hindsight, I think the goal of AI for an HA system is potentially inappropriate as humans are rarely rules-oriented people. Or perhaps better put, any given rule is only applicable for 50% of the time, and the other 50% is a multitude of edge cases that are impossible to accurately predict based on inputs. I'm not sure anything more than VR or gesture control (ie Kinect) will be successful.
 
Even if i'm wrong about AI, we'll always need forms of manual control (VR, gesture, or touchscreen), and HA still isn't going to take off until there's interoperability.
 
Which makes me wonder: If Insteon/UPB are so much better, why don't those manufacturers just build a tiny bridge/1 way controller so they can send commands to zWave devices and make the transition path simpler? If zWave is so much crappier that sounds like an easy customer acquisition play as people won't have to give up their investment.
 
BTW from what I had last read, Insteon/UPB/any PLC weren't an option for me. I don't have 3 wire everywhere, some parts of my house are still 80+ year old knob & tube. Plus the troubleshooting checklist, at least from 2007, was all but a list of stuff I had in my house.  I need a wireless technology, and the WAF won't support RadioRA at $200/switch. At $70/switch, Vizia RF+ zWave with active notification is about as far as I can push it.

 

[pete_c]

Which makes me wonder: If Insteon/UPB are so much better, why don't those manufacturers just build a tiny bridge/1 way controller so they can send commands to zWave devices and make the transition path simpler?

 
Why should they?
 
After all the object of their wares is to move everybody over to their world of automation; that is the way they make money.
 
The multi automation protocol hub is already here. 
 
None that I know of are 1-way communcations.  I do not see a purpose for that in the world of automation.
 
My little Almond + tiny combo router, wireless AP, firewall and automation hub talks X10, Hue, Zigbee and Z-wave 2-way communications bridge.
 
I am running Homeseer today on a tiny Rpi2. 
 
It does talk UPB, Insteon, Z-Wave, Zigbee, X10, Hue, HAI OPII and can manage a small digital IO board.  It talks and does VR if I want it to.
 
HomeGenie, OpenHab, et al can do that too today on a tiny RPi2. 
 

In hindsight, I think the goal of AI for an HA system is potentially inappropriate as humans are rarely rules-oriented people.

 
Yup it's been very easy for folks to get addicted and tethered to their tiny internet devices; no rules involved as its been all voluntary.
 
An nefarious AI created tweet today can start a war with those rarely rules-oriented (they really are rules oriented and do not even know it) people.

 

[zenix]

I need a wireless technology, and the WAF won't support RadioRA at $200/switch. At $70/switch, Vizia RF+ zWave with active notification is about as far as I can push it.

 
What if it is $120/switch? You can also take a look at Caseta, at low price of $54/switch and solid performance, it is hard to beat.
 
http://www.amazon.com/RadioRA-RRD-6CL-WH-CFL-LED-Dimmer/dp/B00JC350NI

 

[drvnbysound]

To me, automation is having a rules-based engine where the house knows what you're going to do so you don't have to do it.

 
I absolutely agree. My issue with this has been that our "schedules" or methods of doing things vary drastically that I've found this hard to implement... Here are some examples:
 
1) When we let our dog out at night, we always turn on the back yard flood lights. I wrote a simple rule, when the back door is opened and it's dark outside, turn the lights on for 5 minutes. This has worked great... easy implementation. If I want to go sit on the back porch, have a drink, and stare at the stars/whatever... well the back lights will be on for the first 5 minutes - unless I manually turn them off.
 
2) Similar to the above, when I open the door from our kitchen, heading toward the garage, I always turn the garage light on if the garage door is closed. Simple rule. However, that light also needs to turn off after X time or it will remain on for hours, or the next day. So, I implemented a time based rule (15 minutes) and tied in some more intelligence with my garage motion. If motion was detected it would reset the time counter. Long story short, I was out in the garage doing some work about a week later and the light kept turning off on me - I'd have to move around to get it to turn back on. I've yet to find a good solution for this other than simply making that timer longer... which means if I'm only in there for 30 seconds, it remains on longer when I come back in - unless I manually turn it off - not automation...
 
3) We have a split floor plan. I've set a rule to turn the living room light on for 10 seconds if motion is detected between 11pm and sunrise. This is primarily for my [toddler] son and wife - a night light if he is heading toward our room, or if she is going to check on him. She loved this. However, I go to bed anywhere between 10:30 and midnight - on any given night. So the nights I'm up later, I could be setting that thing off repeatedly which may wake either of them up. Similarly, on occasion (maybe a 5-10 times a month) I get up before sunrise and I'm triggering the light in the AM.
 
All of that to basically say that I've found that my days vary a lot that it's been difficult for me to implement much in terms of "schedule" based rules, simple because I don't follow a strict schedule. Hell, I might go a month without drinking any coffee, then have it 3 of the next 5 days It's almost impossible to "teach" a system that sort of information.

 

[picta]

@dvrnbysound: you need to add more occupancy and motion sensors to your system and you will be able to automate all of your cases above. But you do need a manual control every now and then, that's given.

 

[jkmonroe]

But 9 years after that first delivering that presentation, I still don't see us being that much closer. My new VRCOP is coming today (from automated outlet, hopefully this one works as the one from amazon never did), I might do some cutesy nit stuff like "trap the code sent to the zWave door lock and use it to disarm the security system" but not sure what else.

 
Good luck.  Attempting this is what solidified my disdain ZWave.  

 

[ano]

+1

I love Zigbee.  It works as expected, and it has field upgradeable firmware.
 
I would love to find a universal, connected, Zigbee bridge - something that can connect the plethora of Zigbee devices with an open API.

I have a growing Zigbee network and it works well.  There is a few second delay to battery operated devices like locks. 120V devices are instant.
 
For an open Zigbee bridge try this: http://smartenit.com/product/harmony-g1/  I think this might work.

 

[drvnbysound]

@dvrnbysound: you need to add more occupancy and motion sensors to your system and you will be able to automate all of your cases above. But you do need a manual control every now and then, that's given.

 
I have had an issue finding a good dry-contact type occupancy sensor. All of the ones I ever find via searches are tied directly to the lighting load, which I don't want. Do you have any recommendations here?
 
Honestly though, I don't know if they would resolve my issues... for example the living room "night light". My motion picks me up fine, so detection isn't the issue... it's that I can't predict a good enough time window when the rule should occur vs. when it shouldn't.
 
The frustrating thing for me is that the need for the manual control seems to be about 50% of the time... so I fight the battle of, should I really spend $xx on controllable lighting that I manually control 50% of the time?

 

[lleo]

@drvnbysound

3) We have a split floor plan. I've set a rule to turn the living room light on for 10 seconds if motion is detected between 11pm and sunrise. This is primarily for my [toddler] son and wife - a night light if he is heading toward our room, or if she is going to check on him. She loved this. However, I go to bed anywhere between 10:30 and midnight - on any given night. So the nights I'm up later, I could be setting that thing off repeatedly which may wake either of them up. Similarly, on occasion (maybe a 5-10 times a month) I get up before sunrise and I'm triggering the light in the AM.

I have many similar rules. Your could enhance your rule to add a button by your bedside to signal your system that it is night now. Otherwise no automation will have the intelligence to know that you are still to go to bed and not just got out to check on the kid. Anything short of mind-reading will not be able to discern this. Big data and the bad guys (google,apple,amazon and the like) may be able to find patterns, see also Nest, but I frankly do not want this. I rather push the button if needed.

 

[IVB]

Why should they?
 
After all the object of their wares is to move everybody over to their world of automation; that is the way they make money.
 
The multi automation protocol hub is already here. 
 
None that I know of are 1-way communcations.  I do not see a purpose for that in the world of automation.
 
My little Almond + tiny combo router, wireless AP, firewall and automation hub talks X10, Hue, Zigbee and Z-wave 2-way communications bridge.
 
I am running Homeseer today on a tiny Rpi2. 
 
It does talk UPB, Insteon, Z-Wave, Zigbee, X10, Hue, HAI OPII and can manage a small digital IO board.  It talks and does VR if I want it to.
 
HomeGenie, OpenHab, et al can do that too today on a tiny RPi2. 
 
 
Yup it's been very easy for folks to get addicted and tethered to their tiny internet devices; no rules involved as its been all voluntary.
 
An nefarious AI created tweet today can start a war with those rarely rules-oriented (they really are rules oriented and do not even know it) people.

 
i'm sure you can do lots of things with an rpi2, but the problem is you have to do lots of work to do lots of things.
 
And why would they build a bridge? Customer acquisition. If someone was told there was a migration path from zWave->Insteon that didn't require throwing away their entire investment, they'd listen.
 
I repeat my wireless networking analogy: If you owned Linksys and it only worked with Linksys cards and linksys printers and linksys USB devices that plugged into your laptop and you had spent $2K or more on it, would you really entertain looking into Netgear if it meant you had to throw it all away and buy another $2K in Netgear?  Now, what if you only had to spend $200 on a netgear router, and over time you could throw out out Linksys stuff and buy netgear? Would you be willing to even consider netgear in the former scenario? And what are the odds of looking into alternative manufacturers for the second scenario?
 
Obviously we know the answer, as the market has dictated that interoperability grows the pie for everyone. Including Linksys, in that example.

 

[drvnbysound]

@drvnbysound


I have many similar rules. Your could enhance your rule to add a button by your bedside to signal your system that it is night now. Otherwise no automation will have the intelligence to know that you are still to go to bed and not just got out to check on the kid. Anything short of mind-reading will not be able to discern this. Big data and the bad guys (google,apple,amazon and the like) may be able to find patterns, see also Nest, but I frankly do not want this. I rather push the button if needed.

 
Yeah, I may have to adjust my behavior to accommodate this... right now they are tied to the "STAY" command, which my wife generally arms when she goes to bed. As above, the issue is that I generally am the one who may stay up later or get up earlier. The issue again, is if the rules are tied to me going to bed and hitting a bedside button, I may be up to 2am trying to meet a deadline for the next day, and if either of them gets up the rule is not yet enabled... there certainly is no easy button!
 
I highly doubt there are any tendencies to my patterns... they are usually side-business related to jobs that I have at the time, which are customer driven.

 

[pete_c]

Quote
I'm sure you can do lots of things with an rpi2, but the problem is you have to do lots of work to do lots of things.

That is not true. Work and lots of things is relative to some one's knowledge base relating to the RPi.

It is also the same relative to what folks consider automation to be whatever it may be.
 

Quote
the market has dictated that interoperability grows the pie

What is dictated and what is done will be different.

It is typically the manner of play that will dictate profit margins for the newbee automation companies.

While interoperability should be dictated and there should be a mechanism to insure compliance of said automation interoperability rules there isn't one today. There are efforts though but their sticks are not big enough yet.

 

[picta]

I have had an issue finding a good dry-contact type occupancy sensor. All of the ones I ever find via searches are tied directly to the lighting load, which I don't want. Do you have any recommendations here?
 
Honestly though, I don't know if they would resolve my issues... for example the living room "night light". My motion picks me up fine, so detection isn't the issue... it's that I can't predict a good enough time window when the rule should occur vs. when it shouldn't.
 
The frustrating thing for me is that the need for the manual control seems to be about 50% of the time... so I fight the battle of, should I really spend $xx on controllable lighting that I manually control 50% of the time?

I have a mix of high quality sensorswitch occupancy sensors and inexpensive motion sensors used for HA only. The secrity system has its own better sensors in critical locations. Asensorswitch is installed in our media room, it senses not only motion but the sound and turns off everything only when no one is in the room. I am not really sure what problem you have with the night light, but say you want it work only sometimes. The easiest way is to use a button that sets a flag which will differentiate between "regimes". Of course you'll need to include this flag in other rules such that it turns off appropriately. The manual button should always be present in any automation system, but over time you can improve your system to have less and less need to push a button, or to make your sceens more complex such that one press will accomplish multiple actions. In our house every light is watched and turned off appropriately, but the mighty "all off" button strategically located near the bed is used a lot.

 

[ChrisCicc]

Which makes me wonder: If Insteon/UPB are so much better, why don't those manufacturers just build a tiny bridge/1 way controller so they can send commands to zWave devices and make the transition path simpler?

 
That's what products like our own are here for... granted some do it better than others. CastleOS and HomeSeer, for instance, fully abstract the protocol from the device. Some, like the ISY, give each protocol its own device type. But they both accomplish the same thing: letting one user interface control multiple protocols. 

 

[batwater]

Posted Today, 02:09 PM

IVB, on 19 May 2015 - 10:40, said:

But 9 years after that first delivering that presentation, I still don't see us being that much closer. My new VRCOP is coming today (from automated outlet, hopefully this one works as the one from amazon never did), I might do some cutesy nit stuff like "trap the code sent to the zWave door lock and use it to disarm the security system" but not sure what else.

 

Good luck.  Attempting this is what solidified my disdain ZWave.  

 
What he said ^^^  I spent considerable time trying to get the logic right to arm / disarm the alarm when the door is locked/unlocked.  Kept getting a causality loop resulting in the alarm arming & disarming repeatedly and the lock locking & unlocking (really sucked to try to get back into the house with that going on.) Of course I was also trying to make it really smart so that if the alarm armed itself and the door was not locked it would automatically be locked.  Separating the events out with z-wave lock/unlock message not being 100% reliable proved to be unreliable enough that I abandon the effort.  Might also have had something to do with the fact that I set the alarm off a bunch of times due to the looping.  
 
Settled on the locks generating an email when unlocked (tells me who unlocked) or when the door is locked (not a who but when) this is great for figuring out if your kid is home or not without having to chip them (something they tend to object to)
 
On this one, with it being the security of the homestead in question, if the behind the scenes automation is not 100% posing any risk of generating false events then its a no go.