I've been into the home automation thing since before forever. There have been many periods of excitement over the years. They come and go. As a result I have to admit to a high level of skepticism about the current period.<p>The companies that generate and are influenced by this excitement tend to make two main errors.<p>1. They don't understand that without an open standard for controls and sensors they are just fooling themselves. No one is going to permanently install something in their house that is sure to be useless in 10 years.<p>2. They don't understand that people don't want to control everything ... they instead want to not <i>have</i> to control everything. Being able to control a light from your phone is not really any better than being able to do it from the light switch. People instead want the light to just do the right thing without having to think about it. They don't want a thermostat controllable from a phone. They want a thermostat that makes it so they are comfortable and not paying any extra money they don't have to. As a result, the home automation people actually want is a much harder problem than what everyone always seems to assume.
Sounds like my experience with Nest last summer. Got the thermostat, can't get it connected to my wireless. Support basically tells me to get bent until I go into full IT guy mode and prove its an issue with the unit. Replacement unit comes, doesn't work at all (first one was at least a nice looking thermostat with no connected features). Third unit comes and it's been fine ever since.<p>For some reason I thought it would be a good idea to get their smoke detector. One day it randomly detects smoke and will not shut the hell up about it even after resetting, replacing batteries, etc. Nest offered to replace it - I just put my old "dumb" detector back up.<p>I will say it's pretty damn embarrassing to buy this stuff and have it fail so spectacularly, at least my roommate wasn't here to see me seriously losing my mind trying to shut up the Protect.
I have a similar problem with Chromecast and Youtube. My friends literally cannot understand why I would rather use the YouTube client built into the TV and control it with the TV remote, rather than cast to chromecast and control with my phone. Controlling the TV with my phone is cumbersome. It takes time to unlock it,to find the app, to send the command and have chromecast react. What if someone rings and I want to pause the video before answering the call? Doable,but more complex than it needs to be. What if I leave the room but someone else wants to pause, skip or whatever, but their phone is not connected to my chromecast? In comparison, TV remote is always in front of my TV. Using it is dead simple,doesn't require pairing,and it's instantaneous. It doesn't leave the room when I go to the kitchen, and I can use it while I am on my phone. I feel like controlling lights with your phone is a similar gimmick. It doesn't make it easier or faster - it's just gimmicky, that's all.
I really don't grok why I should want to use my smartphone to move up and down my wifi connected blinders. Yet that's all I see when all of this is pictured: gimmicks.<p>I want to see a demo of a room alarm clock: with rising light (in brightness and temperature), birds chirping, room getting warmer, music/radio a while later.. and the same for going to bed.
Having each light bulb intelligent seems like overkill, I can see a per circuit setup where you replace the light switch. First off its not going to go bad like a bulb which will be replaced eventually; even LEDs which still die too soon. Really, who came up with this light bulb idea...its just daft.<p>Security motion detectors can be used to double effect here as well, you would naturally have them facing doors so they could activate wireless or through signals across existing wires. All sorts of options.
Automated building systems are hard to get right even in the commercial space where the specifiers, installers, and operators are professionals and end users are given dummy thermostats to make them feel better or just outright locked out of controls and windows are inoperable.<p>Even then it all depends on dedicated circuitry: webs of low voltage wires for power and signaling snaking from control panels in closets. Sure Wifi is capturing some corner cases, but NFPA isn't going to allow its use for fire alarms any time soon. Those systems need to be reliable. Or people die.<p>The smart home won't be based on a wireless bus. Wall warts show that we're already at the point where a standard low voltage power bus makes sense. More wires and a case for surface mounting is the way electrification goes historically.<p>The current approach to smart homes is equivalent to configuring a new network, hooking up a dozen different 80's dot matrix printers, and then writing drivers for them...using a smartphone touchscreen keyboard. Except instead of technical support you get consumer services.<p>Adding a touch screen and 802.11 doesn't make getting a system working or geewhiz configured easier than X10. It just makes it look simpler and cleaner and more likely that the end user will blame themselves for not being able to use a smartphone.
There's a few 'handy' things I'd like to do - for example, I am considering replacing my basic central heating/hot water controller with a Raspberry Pi and having features such as a low-heat, vacation mode where the system knows I am 'x' miles away from coming home (tied to my phone's location) and it will fire up everything again.<p>Although I'd consider doing more (ESP8266 modules with MQTT looks interesting, albeit in its infancy), the proliferation of 'standards' (properitary and otherwise) is offputting. Then there's the standards designed to interface between all the different standards...rinse...repeat.<p>Edit: Spelling - I have a dodgy laptop keyboard with melted keys due to a motherboard failure!! Waiting for Dell to come back again and change it!
There's a lot of cheap crap on the market but there's also some great designs out there.<p>For example, I've got this light control system installed in my home:<p><a href="http://houm.io/" rel="nofollow">http://houm.io/</a><p>Wireless, battery-free buttons that you can put anywhere you like. I keep one of them on my night stand. Rest of the switches look and work just like regular light switches.<p>Except the switches are connected to the system, so I can activate pre-defined scenes with one click of a button. No need to use a smartphone. The scenes are configured through a multi-platform web client.<p>The client, by the way, uses a regular REST API so I can control my lights with cURL too. Over the internet :)<p>It's more expensive, but it just works.
Maybe I'm just not getting it, but why would I want a "Smart Home"? It seems to be a little like a Smart TV, everyone wants more features, but really they're newer going to use the built in voice control or install any apps. We need a few basic feature and the rest just sits their, unused, contributing to the complexity of the device.<p>I have yet to see an intelligent house feature I would like. I understand that some want a better climate control interface, like the Nest. But I have yet to see a house that have central climate/heating control, and I don't miss it in my own home.<p>Is it really such an inconvenience to get up to turn on the light or turn the blinders?
I have had auto blinders, lights, and day alarm (courtesy of an old projector) for years. It's all controlled by some random netbook I got for $30, picaxe micros, and xbees (the xbees are admittedly overkill).<p>This is the sort of thing you build, rather than buy, since it will only ever fit the person who designed it.
Not sure I agree with the article's title. He's basically bad mouthing the entire smart home industry when this really seems like more of a Quirky/GE/Wink mess up. Literally every single product he tested is by them.<p>I actually LOVE all of the stuff in my smart home! All I have is 6 hue bulbs, a Nest, and an Amazon Echo that I've programmed to control everything with my voice with the help of a Raspberry Pi.<p>Nothing feels better than walking in after a long day and saying:<p>"Alexa, turn on all the lights, color blue, brightness 9"<p>"Alexa set nest to 72"<p>Then sitting down on the couch and saying:<p>"Alexa play some jazz music"<p><i>pours bourbon</i><p>If you're interested in this, check out my Alexa Home project alexaho.me or the YouTube video below) . I'm turning Amazon Echo into a smart home controller and people seem to be pretty jazzed about it. Frankly, if we can get stuff to work like this where it doesn't rely on your smartphone, people will use it a lot more. I also have a small switch next to my bed that turns on and off the overhead light and desk lamp in my room for when I don't have my phone on me.<p>I think success in the home automation biz is going to be all about untying it from your smartphone...<p><a href="https://www.youtube.com/watch?v=9AmxiGVBekE" rel="nofollow">https://www.youtube.com/watch?v=9AmxiGVBekE</a>
I know our generation is the one longing for the concept of a smart-home that we see in sci-fi movies, but until we reach a stage where we can actually give contextual voice commands that our "home" can understand I'm not convinced this is worth spending a dime on. Not even in the "to promote research in that direction" kind of way, because all I am seeing now are people taking advantage of our high dollar whims and are missing hard.
I feel like there's some kind of rush to become the box/service that runs your smart home. <i>None</i> of these vertically integrated solutions is going to cover all of the use-cases or user desires, yet it seems that everyone want to be the one that sells you their proprietary (and bug-ridden) box.<p>It would be much easier if most of this stuff could be shipped as software that you can install (somehow) into a box of your choosing and pull together components for your home as you see fit. This field isn't ready for consumers per se but those with a more hacky mindset would probably uncover the useful bits, which could later be picked up as integrated solutions. The only thing I've come across that's like this is openHAB (<a href="http://www.openhab.org" rel="nofollow">http://www.openhab.org</a>)
My main issue with my wink is that it has to go all the way across the internet to connect to the box 10ft away from me. A much "smarter" system would simply connect directly over my home wifi and avoid most of the lag and connectivity issues.<p>I understand WHY they did it this way (because they don't want to support an open device/protocol for interacting with the hub...) but it seems really short-sighted in terms of the user experience.<p>I've already rooted my wink hub and fully intend to implement this myself just to reduce the annoyance.
Homeboy cameras are the one thing in my setup that work perfectly. I wish more products could be so well integrated with the cloud and the phone. I have them set up to send pictures to Dropbox via IFTT, and it all works perfectly, recording only when there's motion, and only when we are not home (it uses your cell phone location).<p>Tried a Quirky Spotter, and it does detect things like motion, but it doesn't know to send just one push notification; instead it sends a continuous stream of them while motion is happening, which is irksome.
This article made me think about these idea for a bit. I thought of something that may not be very practical right now, but could be a way forward for these things. I think what a really good home automation system ought to do is a bigger version of what the Nest does.<p>All of the light switches, blinds, alarms, climate control, music, TVs, etc could be manually controlled at first, but there's a system that watches how you change all of them and coordinates it with your location and the time. It could look for patterns of doing several things in a row or doing them at the same time, and either offer to or just start doing them all together for you. Like you normally turn on lights 1, 2, and 3 when you walk in the door from work, so it notices and starts turning them on for you just from opening the door. Or you normally turn off lights 2 and 3 and turn on the DVD player within 5 minutes of each other, so it starts just turning off those lights when you turn on the DVD player.<p>Basically, something that notices what you already do and tries to do it for you, without you having to program things in with a smartphone or something. I bet normal people would love that. As long as it wasn't laggy and buggy, of course, which could be easier said than done.
There are house-integrated things that are cool, but they rarely require integration. CO2 and humidity-controlled ventilation is really nice, as is whole-house sound systems. But why should they be connected to each other?<p>Maybe if you used some machine learning algorithms and made the house actually intelligent, so it could anticipate what you wanted. Other than that I don't really see the use.
I think a big part of the problem is that integration is hard, and that the companies more or less resist it (I guess because they think keeping people in their garden is a good strategy).<p>It's also easy to oversell (to yourself) the value of a room having a movie mode. Most smart home features are 'nice to have', not worth a lot of messing around and cash.
There's a general pattern of trying to dumb things down to a simple, inflexible interface, too often with half-baked "AI" bolted on. Good products still need attentive human supervision.<p>I installed Insteon motion detectors and webcams after a robbery, but the included software was such undependable and inflexible garbage that I replaced it all with a simple Misterhouse-based Perl script that sends texts via email.<p>If I see another tech product ad aimed at millenials featuring bright easter colors and indiepop music pitched by Steve Jobs wannabes who are unable to get any angry nerds to make their products actually work, I might snap.<p>By the way, the other HN article about malfunctioning Nest smoke detectors is a hilarious accompaniment to this. ( <a href="https://news.ycombinator.com/item?id=9048110" rel="nofollow">https://news.ycombinator.com/item?id=9048110</a> )
So-called "smart home" devices are certainly new and imperfect but the story told at the beginning of the article is also not necessarily representative.<p>He describes months of setup, unlocking his phone, scrolling through apps, finding the right screen, drilling through menus, and then not having it work. In my limited experience with similar setups (a few Philips Hue bulbs, IP camera, and a Nest thermostat) I haven't really run into anything that obnoxious.<p>When I want to kick on the "mood lighting" for a movie or a party or whatever, I just wake my phone and tap one of several presets on the homescreen widget. Setting up those presets took maybe 10 minutes and I can always mess with the lamps manually or create more if I feel like it. Otherwise the handful of IFTTT recipes I set up and geofencing keeps the lights under control the majority of the time without any input from me.<p>Nest is similar. It's reasonably good at learning my patterns and if I want to tweak the temp or change something, I just open an app and the image on the screen matches the display of the physical thermostat.<p>These are certainly early implementations of connected/smart devices but compared to the state of "professional" integrated systems, they're just fine with me. Just read up on Crestron programming if you want to learn how complicated the enterprise solutions can be. And if you don't want to spend months learning their wacky programming tools and procedures you can spend thousands on hiring certified techs to do it for you.<p>The thing I'm most excited about is for more standardized and interoperable devices to start overtaking the pricey "pro" options for at least the lower level stuff in enterprise and education. As it stands now, a lot of the Crestron stuff is not much more than a glorified universal remote. The faster the consumer/DIY stuff improves, the sooner it will be able to replace the ridiculous "pro" stuff that's been the only option for years.
I am now on my 4 controller "Smartthings (ST)" after dumping Nexia my 1st and for many reasons such as compatibility and monthly costs et cetera, then Micasa Verde whcih was okay but not the creates their tech support sucks and more incompatibility issues, then Indigo6 which was okay for a year but they drove me nuts refusing to speak to my doors locks (schlarge) and had to use vera bridge which couldn't keep proper track of the state of the doors.<p>All the while I was researching a stand alone controller backed by a big company like google, apple or samsung. Then I found ST backed by Samsung.<p>So here we are. migrated and all. Learning how to deal with MODES OMG. No variables I can set to do things.
When you use a smart phone as a remote control, it's just a slow and inconvenient remote control device. If the lights don't dim by voice command, it should be by your regular smart remote or the switch.<p>"I unlocked my phone. I found the right home screen. I opened the Wink app. I navigated to the Lights section. I toggled over to the sets of light bulbs that I'd painstakingly grouped and labeled. I tapped "Living Room"—this was it—and the icon went from bright to dark. (Okay, so that was like six taps.)"
Am I the only confused by the crazy delay mentioned in the article from the "Robots"? The processing can't be all that intensive, and I can't imagine a setup where communicating simple on/off commands to specific devices should be so costly as tens of seconds. I would understand a general delay from sensing the triggering event (i.e. opening the door), but that hardly explains the delay in between all the lights as shown in the gif.
The best 'consumer facing' demos of these involve the "I'm leaving home" or "I'm coming home" activities. Aggregate functions, think IFTTT.<p>For much of these devices today, the experience is so fragmented that only a tech savvy customer can figure them out.<p>Also, yes, it's not helpful when I have to take out my phone, put in the password, open the app, and click 'off' when I could just get up and press the switch much faster.
Home automation is not something you can install yourself, unless you are a full stack programmer, and with full stack I mean also PCB's and micro-controllers.<p>Surprisingly there's a lot of cheap consumer stuff out there, witch can easily be modified to do what you want. Some of them are "unhackable" though.<p>The trick is to make everything automated. And use a high level language like JS to make it smart. Yes, I write my home automation in JavaScript :P
It seems like the vast majority of the issues with the current generation of products on the market has to do with speed & reliability; thus hardwired is the only way to guarantee those 2 work all the time. Also by going the hardwired route, the sensors and controls will be vendor neutral; thus ensuring that it will be future proof.<p>After doing a bunch of research the conclusion I came to was Opto-isolated relays [1] are the way to go for lighting control [2]. This allows for simple logic boards to control line voltage things without the risk of the line voltage killing your logic board. Best way to do this style of home automation would be in new construction, as retrofitting it would be hard to do if the house is more than 1 story tall as you need access above and below. Would be labor intensive as well. Will also allow for manual control because you'd put a current sensor on the wire so the hardware can tell when the circuit is on/off.<p>I would also want to put in some thermal cameras, as motion sensors are not as useful when your sitting on the couch, to tell when you enter/exit a room for auto lighting control. Put temperature sensors in every room and have motorized dampers & air flow sensor allowing for the control of the temperature in every room. Now your lighting and hvac can be manually or automatically controlled. Wire all the sensors and controls to a hardware/software brain that can be upgraded at any time. If that hardware/software brain is disconnected, all the light switches will still work and the normal hvac system will still work; just less efficiently.<p>Have the thermostat be controllable from the panel when the house is occupied (thermal camera) and auto controllable when vacant.<p>All of the above is possible right now; it just would cost a lot of money. Every relay/current sensor is $80 bucks; every damper control is $100; every thermal camera is $500; add it all together and you got one very expensive home automation system. But it will work right now without a lot of user frustration.<p>[1] <a href="http://en.wikipedia.org/wiki/Opto-isolator" rel="nofollow">http://en.wikipedia.org/wiki/Opto-isolator</a>
[2] <a href="http://diy.stackexchange.com/questions/46601/home-light-automation-emulate-4-way-switch-with-relay/46610#46610" rel="nofollow">http://diy.stackexchange.com/questions/46601/home-light-auto...</a>
Home security too. Finding an HD, outdoor camera system that has smartphone and web functionality is very difficult. All the NVRs are horribly insecure. It's strange that a (physical) security product completely neglects software security. The more modern companies seem focused on indoor cameras with only one or two variants and not much by ways of optical zoom lenses.
One comparison to this is the big push to get in car navigation and smarts developed by the tech sector. There is a reason radios and stuff work year after year through heat and cold and the like.
they may not be pretty, but it (usually) does what its suppose to even after years of use.
I think part of the problem is that frequently/mostly, I want a dedicated input device like a switch or a dial, not my iPhone to control stuff like lights.<p>Nest has an app but primary interaction is with dedicated UI widgets.
What is the power limit of POE? I don't see a reason to try to dim LEDs on AC or a reason to keep converting DC/AC/DC/AC so often. Wouldn't POE or a similar standard be better?