Blue sky discussion (moved from elsewhere) - what a mature smart home might comprise

General Discussion
1

At the behest of another poster, I moved this from another thread so it would not interfere with people finding information. Please feel free to share your thoughts.

This post is kinda scary. Not in a bad way, just in terms of a more general user acceptance. As I along with many others have observed, the smart home - while long dreamed - remains in its infancy from a concept/implementation standpoint.

For the smart home to really begin to gain general acceptance, it needs to NOT have to do “CORE pistons” or require a complex half-hour project to simply AskAlexa “is my garage door open?”.

The smart home, to gain real acceptance, will have to

A) distinguish one voice from another, for both practical and security reasons. You do NOT want a random stranger to be able to shout “Alexa, unlock the door” and have it work.

B) be easy for a user to configure by voice: when John says “Alexa, turn on movie night” Alexa would reply “john, tell me what comprises movie night”. HE then says “turn on the screen, set the input to bluray, turn on surround sound, make the overhead lights very dim and blue and turn off all the lamps in the room” and the smart home configures it. Then when Jane says it for the first time alexa asks her “jane, John already set up a movie night. Would you like your configuration to mirror his?” She might reply “yes, but make the overhead lights dimmer and a darker shade of blue, and leave the table lamp on but dim. And make it louder” and it would do that.

C) Local processing. Given the falling price of computing power, we would want to rely on the web - with its potential outages - as little as possible. The smarts behind your local processor might update automatically, but they would still be local.

There’s more, but that’s a start. Until then, it remains mostly a hobby for geeks and a nightmare for their wives :smiley:

3 Likes
Will SmartThings have a future? (2019)
2

Interesting topic!

First, my usual note (Hi, Mom!):

There are a number of community members who are female, some with wives and some with husbands. And of course some single with a bunch of roommates that they provide the tech-support for. My mom was a senior programmer for AT&T and then taught computer science at the college level. My dad has no technical skills whatsoever. Gender doesn’t define tech status. Just sayin’…

Now on to the discussion!

I think we’ll see widespread acceptance with a much simpler feature set than the one you describe. In fact, I think we’ve already seen it with both Amazon Echo and Philips Hue. I don’t think a lowcost HA system has to be as clever as you suggest–it just has to be easy to set up and it has to run reliably.

We know what the initial use cases will be for many people:

  1. turn on the lights at sunset and off again when I want

  2. turn on the lights when I get home after dark

  3. turn on a group of lights that I specify and that may be in more than one room both with a handheld device and with one voice command

  4. turn on a light when someone walks into the room and off again after they’ve left

  5. unlock the door when I get home

  6. unlock the door remotely if I want

Honestly, just that much is probably enough for widespread acceptance–if it’s a set and forget system. Of course, you can get that much now with HomeKit, so we’ll see where we’re at in a year. :wink:

The big question to me is still why doesn’t SmartThings provide prebuilt receipes for these 5 use cases? Why make the customer do so much of the set up work?

2 Likes
3

I agree that HA is in it’s infancy still…

I don’t necessarily agree that rule engine functionality is not part of what HA should be.

Even if it’s transparent and part of a AI in the background it’s still building a rule engine - you walk in room during these hours, turn on these lights… that’s a rule engine.

As far as the voice stuff… that’s very individual. I don’t care of it, not sure I ever will.

4

Interesting five - no six - use cases. I see what you mean bout ‘canned’ recipes, but the moment you go there you have to start customizing. Because there’s no way for a packaged product to know WHICH Lights you want to include in your #3 case. Heck, you probably don’t even have any ‘smart’ lights in your home at first! How does a vendor make that happen? There we immediately segue to my suggestion of voice configuration: Alexa, please create a group of lights called “main lights” that go on and off when I say so, and include the living room and dining room lights in it".

As for the recognition of only ‘authorized’ voices for certain functions: my dad, who is not stupid but not quite up-to-date with some tech, when I showed him the voice-controlled home he immediately asked “does it recognize different voices?” So clearly, that is a security feature that would go a long way to making an average consumer feel more at ease about setting up something like this for themselves.

5

Sorry if I wasn’t clear. The sixth use case is obvious to set up in SmartThings. Toggle the lock. You can do it from home you can do it from the office 10 miles away.

But the other five use cases are not obvious when you first get smartthings. And they should be.

Again, sorry if I wasn’t clear, but I didn’t mean those should be the only use cases. I think there should be more, I think if there’s an AI that smart enough to do the voice conversation that’s great, but I think a regular rules engine is fine in the meantime. The SmartRules app is A good example – – only one level deeper in complexity than a standard SmartThings routine, but that opens up a lot. And it’s a much more intuitive layout.

http://smartrulesapp.com

IMG_1603.PNG640×1136 163 KB

Again, I think something with the complexity of core is great, and it should be available for the people who want to use it. It just shouldn’t be the only way of doing some of these obvious things. :sunglasses:

One that I left out of my basic use cases is the management of RGBW lights, again in user-defined groups. Or at least the ability to very easily execute a Phillips hue scene that was set up in the Phillips hue app.

I think voice is great, I use it all the time. But I have friends with different physical abilities. Some of them can’t vocalize. Some of them can’t hear. Some of them can vocalize and hear just fine, but cognitively they just find it easier to process visual text and icons than to process auditory information. Alternative control methods have always made the most sense to me, even before my own illness. I just don’t think there’s one best method that fits everybody. One of the things automation ought to be able to do is to provide a wider choice than mechanical control methods.

JMO

6

I think the biggest ‘behind the scenes’ piece that will help drive any of this forward is the adoption of locally accessible APIs from every vendor with a simple but secure authorization model. The more industry standard the better but that’s more of a dream scenario. The way it stands today I can’t access SmartThings directly without another integration layer, my plugs have a local API, my locks have no API and required workarounds (beyond most consumers and not scaleable beyond a one off consumer such as myself), my thermostat has a cloud only API with no local option and authentication that constantly has issues and the SmartThings hub has no real API and requires something like CoRE to be able to make simple web calls into an action group. My cameras have no API and require a separate WiFi hub, Hue requires a hub, they all have a separate application + the ST hub … I’ve got a rack full of equipment. It’s all over the place and I’ve done only a fraction of what a lot of people have…

20 or so switches, a couple of locks, thermostat, plugs, Hue and few Echos and I’ve got a web of shell scripts making web calls, local web servers, Amazon Lamdas, services running in the cloud and virtual machines just to make it all work the way I want…

Open and local APIs everywhere, limit the hardware required and establish some kind of industry(ish) wide publish and subscribe service bus similar to MQTT and the rest will start to fall into place.

Just my 6 cents… :slight_smile:

7

There are working groups to provide IP enabled Zwave and Zigbee.

If that ever happens, a restful API could be made on a open compute system that could talk to and control those devices.

The would then handle all the things you could cover under those Z networks - but that doesn’t cover all the devices you/we want these days.

The problem is the providers are often guarded and closed and don’t necessarily want that open unfettered access for various reasons.

8

Definitely agree. Companies have a vested interest in keeping you in their cloud environment / stickiness and various working groups seem to move at a snails pace when you throw multiple vendors with, to your point, vested alternative interests. Look at the IEEE just as an example… IT titans are actually coming together to bypass them for various standards simply because of how slow they move.

Not sure the best answer but I digress…

As far as what a mature smart home might look like… I think you absolutely need a natural language interface but it should be secondary to action. Meaning the home, the car, various sensors should be smart enough to know what I need and when. Lights, temperature, directions, reminders, etc… these should all happen automatically without me having to put decision cycles into them and, which is missing today, seamlessly. I think of it as the wife or husband who knows to remind the kids to feed the dogs, who knows when you’re out of milk, who knows how you like your bath (that one may be stretching it but you get the point), etc… my wife doesn’t have to remind me to feed the dogs or order more milk or grab wine from the store because it’s Thursday. I just do. These are decision items that a good engine could off-load.

The natural language should be more of an assistant. I notice you’re clearly leaving the house and don’t have an umbrella… it’s going to rain. You’re looking for your keys I see… they are on the counter, let me ring your cell phone since you’re stomping around the house and cursing that you can’t find it, etc… you don’t have to tell Alexa to call your phone or ask her for the weather. She knows as you’re walking out the door.

Presence and vital sensors would also be part of a mature smart home in my mind. I wouldn’t have to worry about aging parents falling or having a heart attack when I live four hours away if an intelligent system could clearly know that a fall has occurred, vital signs are indicating a problem or know definitely that medication hasn’t been taken and to alert me if they refused / needed help, etc…

That being said the privacy and creepiness factor is a whole other story…

2 Likes
9

All this sounds nice but to what end? What is the ultimate goal here? To automate our lives into monotony? To have our houses, cars, whatever be smart enough to do what we want before we know we want it? To perpetuate the increasing laziness and decreasing self reliance that has taken an ever strengthening grip on modern society?

That is all, existential crisis over.

1 Like
10

Not for me.

But to free ourselves from a particular hub or locked in eco system, an open api for an would be nice.

11

I honestly don’t care if it’s an open system or not. I just want an affordable reliable system that meets some pretty simple use cases.

For me, affordable means $300-$500 per room. And I don’t have to do everything in every room. And I don’t have to do every room in the house.

I could get everything I wanted today from control4 now that they’ve added echo, I just can’t afford it.

But I really don’t look at these systems in terms of features. I look at them in terms of what problems I want to solve. So I’m less worried about future proofing then some people are, because I’m trying to solve very practical problems now. :sunglasses:

I think of home automation the way I think of cell service. There’s a certain amount of money I’m willing to spend per month, and as long as I don’t go over that, I’m fine with replacing devices every three years if needed. So, sure, 10 years from now they will be home automation systems that can track heartbeats as people move around the house. I’ll fold those in as they fit my regular equipment replacement cycle.

But reliability is still my number one requirement.

1 Like
12

Open api is easily achievable… as long as we don’t rely on the corporations to provide , at least in a timely manner

13

Well that’s part of the problem isn’t it. There will be as many different definitions of “smart home” as there are people.
Let’s take the idea of a smart refrigerator. Some would contend that it should know when your ‘staple’ foods are running out, and re-order them for you automatically as they deplete. But what if I decide I’m tired of Hellmanns mayo, and want to try the store brand this time? There’s the Hellmanns, already sitting in my pantry. Absent some sort of intervention, I’m locked into Hellmanns forever. That’s not too smart!

And then there’s my fave conversational smarthome scenario: you like the lights a certain way when you enter a room. So you automate them to do that each and every time you enter. What about your wife’s preferences? Whose preferences get automatic priority? How about when her mom comes to visit? Does mom get to have lights the way she wants, ever? Again, that could easily led to some not-very-smart things… if not to outright war within the family.

14

Mine do of course…

And don’t tell her I said that or pay any attention to this post self destructing if she logs in

2 Likes
15

I don’t see this as a smart home issue. Families have this even with manually controlled devices, they’re familiar with the kinds of compromises that have to be made. Several studies indicate that most women do prefer a room to be warmer than most men, and conventional married couples have to deal with that.

A truly smart home would have multiple scenario options: one when only Mike is in the room, one when only Lisa in the room, one when both are home. Which is just mirroring the same thing that they would do with manual controls. And eventually multiple iterations of those. But that’s all super futuristic. It’s not going to be necessary for mass-market acceptance , which is where this thread started. Obviously nest faces exactly the same issue now. Families make compromises on everything from the brightness of the lights to the setting of the thermostat. :sunglasses::princess:t2::dog::smile_cat:

16

What? I’ve never had an argument with anyone in my household about the thermostat temperature… does that actually happen? :wink:

Good point but as far as far as monotony @Crussell I view it a bit differently. Smart enough to know what we want before we want it? Absolutely. I / we all make a million mundane decisions a day. Turning off the lights, adjusting the water temperature, forgetting an umbrella and walking into that customer meeting soaked, all decisions that if I could relieve myself of all the better. Granted these are just fun / BS examples for illustration but… I don’t see it as contributing to laziness at all. I see it as a tool giving me more time for family, the gym, get that big promotion when I don’t walk into a customer soaking wet. My daily life runs 18 hours a day and anything I can do to be more efficient and better prepared I’m all about. If I can sleep better the other 6 knowing distant relatives were being watched over that interests me.

That being said could the overall contribution of a fully automated house / lifestyle lead to complacency and laziness? More time to have that mobile device in-hand avoiding human contact? Making it more convenient to not call that distant relative? No question. I suspect people who fall into that category are already lazy and complacent anyway, though.

In full disclosure I probably wouldn’t use that extra time for the gym… :wink:

1 Like
17

I absolutely and totally disagree about disagreeing about agreeing about me disagreeing to disagree with your agreeable statement.

What was I saying? What I said was mainly sarcasm, but something to think about. I spent 12 years in school learning all the ins and outs of ai, automation, electrical/mechanical engineering, reverse engineering, etc so obviously I’m very invested in IoT. I have a ton of automations set up and my own cloud for some of the stuff I work on so I’m the last person who should say anything :innocent:

1 Like