Voice Activated Home Automation
- Thread starter heffneil
- Start date
The walls are drawn using Visio. But all of the controls, windows, doors, thermostats, card readers, and such, are custom controls. They respond to events from the HAI, and can trigger events. Doors for example turn from green-hatched to red-zigzag when they are opened...You don't see in the video, but the "buttons" that say "burglary" and "fire alarm" and such are also custom controls. They light up and flash when they are in alarm. It looks great. If multiple alarms are occurring, the flashing controls are synchronized. The plan is to put this on a touch screen embedded in the hallway in the house. This happens to be windows forms, not wpf, but no particular reason.
jmorris644
Active Member
Take a look at iRule. I have been able to do some great things with the combination of iRule, HaikuHelper, a web server, and of course, HAI.
Joe
elvisimprsntr
Active Member
SiriProxy
Nice!
ChrisCicc
Active Member
Hey Folks, just saw this dedicated VR thread, so please forgive the post here in addition to the thread I started. I'm the founder of CastleOS, a new home automation software suite that features whole-house voice control by using the Microsoft Kinect! This is not a hobby project, it's a packaged product that is currently in open beta testing.
The CastleOS voice control interface works just like the OP described, it listens persistently in the background of a room 24/7, waiting for someone to say something like "computer, turn up the heat". It also speaks back to you, for instance when you ask "computer, what's the forecast for Sunday". There's no device to activate or press, it's just there, listening, and it's remarkably accurate.
I can't post links, but we have demo videos and screenshots posted on our website: CastleOS.com
I'll reply here to any questions as well!
The CastleOS voice control interface works just like the OP described, it listens persistently in the background of a room 24/7, waiting for someone to say something like "computer, turn up the heat". It also speaks back to you, for instance when you ask "computer, what's the forecast for Sunday". There's no device to activate or press, it's just there, listening, and it's remarkably accurate.
I can't post links, but we have demo videos and screenshots posted on our website: CastleOS.com
I'll reply here to any questions as well!
BraveSirRobbin
Moderator
I'll believe it when I see it. IMO, voice recognition is just not ready for prime time, especially trying to handle the background of a busy household.
Also, ChrisCicc, please list your affiliation with CastleOS in your signature, as required for vendors posting here.
ChrisCicc
Active Member
Just updated the signature. I did try to do that at first, but I couldn't access my profile settings while the account was probationary apparently.
As far as voice control being ready for prime time, I fully expect CastleOS will be the product that changes your mind. We have one demo video posted now on our website, and four more now under post-production coming soon. If you have a Kinect and Insteon at home already, it won't take long to set up CastleOS just to demo the voice control.
As far as your concerns about a loud and busy household, while sometimes CastleOS will not be able to isolate your voice over the background noise, it will also almost never get a false positive from background noise. The false positive rate is so low I don't even quote it in percent, I instead put it in relative terms: running 24/7 in a living room may result in a false positive about once every three months. Personally, my girlfriend and I leave our condo by saying "Computer all lights off" literally every single time we leave, and our front door is over 25 feet away from the nearest Kinect. It really works!
BraveSirRobbin
Moderator
Curious, do you have kids and a dog?
ChrisCicc
Active Member
Haha, no, but one of my main test homes does have two (somewhat loud) dogs. Also, I work from home with loud TV/music on in the background consistently. It doesn't matter the type of noise, only the level of noise. On a technical level, the speech to be recognized must be 20db louder than whatever the background noise is.
I've expanded on the excellent SiriProxy work of elvisimprsntr (and others) and wrote a direct-to-Elk M1 plug-in for the Raspberry Pi. It supports lighting, tasks, thermostats, temp sensors etc. I'm happy to share, just don't know how to push it up to github.
Mike
wuench
Senior Member
While I share your skepticism, the Kinect on my XBox does a pretty good job. I was pretty amazed when I first got it. It's not perfect, but it does work throughout my 30x30 home theater room even with background noise. The Xbox does have the benefit of knowing what it is playing so it can cancel out the audio. I would say it is 75% with a movie playing loudly. Better with no movie playing.
I don't see how VR will ever meet people's expectations, I mean if a person can't hear/understand someone across the room at a party VR is not going to do it without some yet to be invented technology, not just software. But intelligent mic's like kinect are a step in the right direction....
elvisimprsntr
Active Member
mghan said:
Awesome! Yes please share on GH. You can post to GH using the old school command line method, or you can simply download the native app for your OS distribution which makes it 1000X easier to mange your GH repos. The easiest way is to clone an existing repo and tweak from there.
http://windows.github.com/help.html
http://mac.github.com/help.html
Feel free to plagurize from one of my repos
https://github.com/elvisimprsntr
I also encourage you to edit Plamoni's SiriProxy plugins page to add links to your repo and demo videos.
https://github.com/plamoni/SiriProxy/wiki/Plugins
Anything to help encourage the use of, in my opinion, the most cost effective and practical solution for HA voice control. Show me a solution that costs less than $35 USD, uses an always by your side device with multiple noise canncelling microphones, has the ability to provide both voice and visual feedback in the palm of you hand, from anywhere in the world, without a recurring monthly extortion fee, and I might change my opinion.
P.S. I posted a new RPi SD card image with the latest Ruby and a few other improvements. https://sourceforge.net/p/siriproxyrpi/news/2013/01/release-2013-01-24/
ChrisCicc
Active Member
It will be hard to compete with the price of SiriProxy, since it will always be free 
Also, according to the SiriProxy website, it will only work on a home network, not outside the home.
As soon as Apple releases an official Siri API, we'll add that voice support to our existing Kinect voice control system.
Also, according to the SiriProxy website, it will only work on a home network, not outside the home.
As soon as Apple releases an official Siri API, we'll add that voice support to our existing Kinect voice control system.
elvisimprsntr
Active Member
Actually you can make SiriProxy work remotely with simple VPN. I do it all the time.ChrisCicc said:
Similar threads
