One of the trickiest things about making parts of your home ‘smart’ can be having them recognize and work well for multiple people. It is great to have your thermostat save you money while you are away or your lights automatically turn off or your garage door to close when you are not at home. The problems start when you head out to grab some last minute groceries and your spouse is suddenly left in the dark, with no heating or cooling and doors, blinds, etc. creepily closing themselves up and locking them in. This is when the ‘smart’ home becomes pretty dumb. There are probably more smart homes that have been ripped out or disabled because of this than because they don’t work the way they were designed.
Design is an important factor when you are selecting smart devices for your home. Many of the market leading home automation products either differentiate themselves by offering advanced methods of personalizing and recognizing multiple people in a space. For example: Nest uses multiple sensors including motion and noise sensors in their thermostats to determine if a space is occupied. The Nest thermostat can also use your smartphone as an additional indicator to track when you are home or away and can track up to 10 people as part of a family. Taking this to the next level Nest also integrates with other devices in their ecosystem including the Nest smoke detectors and cameras and other “Works with Nest” partner products. The sensors in each device work together to trigger and activate or deactivate services.
You don’t need to use products from only one vendor but that can make it much easier to design and removes a lot of complexity. We will talk later in this post about some of the different ways to build a well balanced smart home ecosystem with more advanced integration tools. If you want to jump right in at the deep end skip ahead to here: The Advanced Class.
The next sections cover the basics of shared spaces and some of the most fundamental considerations. We provide our recommendations in each section to help avoid common pitfalls.
It is easy to get carried away once you get started with home automation. Some of this stuff is really cool and almost too easy to install. So the bedrock of any really good smart home design (and particularly important for shared spaces) is good planning.
Probably the most basic and often overlooked question is “Who”. Who will be interacting with this system. While it seems easy on the surface it can often the most difficult to answer. It usually starts off simple: Will the smart home recognize me? Well, of course. My spouse? Yes. This is easy! Kids? Yeppers. Anyone else living in the house – sure. How about extended family? People who look after things while you are away? Friends? Cleaners? Dog walker? Wait, do pets count as a person? Gagh!! This is getting out of control.
This can get daunting and you will probably quickly decide that you don’t care enough to worry about someone who is not around regularly. We agree. Despite that, it is still a good idea to think about all the people who may be in your space and how they might impact or interact with your home automation system. Believe me; thinking about this now can save you many headaches later. No one likes headaches.
Our recommendations when you think about who will be in the space are pretty common sense:
- Follow Occam’s razor or the KISS principle and keep it as simple as you can.
- Keep the number of people specifically monitored to as few as possible
- Minimize the number of sensors and use general occupancy sensors as much as possible
- Ensure everyone who uses the space is aware the system is identifying them and where necessary make sure they know what the system is automating for them
It can be very tempting to try to automate everything. I mean, you installed that thermostat easily enough and it saves you money and makes you more comfortable. How can you beat that? Let’s try some smart lights. They change colours, use less energy, light up when you need them to and turn themselves off when you’re not around, you can configure them to alert you when something has happened, use them to help keep your house safe while you are away and they are super easy to install. Wow. Like eating that first bite of cheesecake – delicious! Where have you been all my life? Home automation is good stuff.
Then, like devouring an entire cake for breakfast, you wake up and realize that even your crockpot and the eggs in your fridge are connected to your smart home. What happened? Why am I suddenly a slave to my stuff? Like anything in life deciding what you automate is a balance and there is definitely a point you will hit where there is too much of a good thing. The problem for shared spaces is that the tipping point for ‘too much’ is rarely the same (or even close!) for all of the occupants.
- Consider general occupancy automation (i.e. things that will impact everyone in the home like heating/cooling and lights) very carefully. Discuss changes you want to make with everyone impacted. Try to get others excited about what you want to do and consider how it will impact them and preemptively address their concerns.
- Try it for yourself first. Wherever possible zone off a section of the space and work the bugs out of the system before you blow up someone else’s world. This isn’t always possible or necessary. For example, you could try putting a smart thermostat in a room with electric baseboard heat for a while but you probably wouldn’t get a good idea of how it will work. On the other hand a good example might be setting up a voice controlled home automation system in your bedroom or office and use it there before moving it to a family space.
- Add stuff in stages. Best if you can add new devices or systems one thing at a time. “But this stuff all works together” you say. “I want to get to the full integrated power of the smart home ecosystem” you say. We encourage and applaud your enthusiasm; just remember the cake. There are obvious exceptions to this rule: if everything you are adding is from one vendor you are probably okay or if you need to install a hub and a siren and a door sensor for the security system you are building; that just makes sense.
- Think carefully about how the pieces will (or will not) work together. You may be able to hack things so that anything can work together but just because it is possible doesn’t mean it is a good idea. Look for platforms like ZWave, Bluetooth or Zigbee to ensure that a common communication protocol exists between products from different vendors. Simple systems are more reliable and tend to be less frustrating for others in the home who maybe didn’t even want them there in the first place.
- Think about how many devices you might ultimately have. Will you need a different app, controller, etc. for each or groups of them? You may be able to centralize control with a control hub (like Wink, Vera or HomeSeer) or meshing service (like IFTTT, Zapier or Flow). Regardless of how amazing your home automation kung-fu is you need to remember you are designing the system for the weakest link. If the luddite in your life can’t figure out the system something is bound to break. Is that what you want?
- If you do use a control hub select the hub and devices with extra care. Many devices/services offer limited functionality outside of their native control interface.
- Think carefully about the touch points. How the system is controlled will significantly impact how it is used. If everything you do requires opening up an app on a tablet or phone that extra step may well be the death knell for the system. The best solutions either mimic their dumb counterparts or add substantial functionality.
Once you have a good idea of the automation you want, the next step is to consider how or even if the solution will work for multiple people in your home. If you are not building or renovating the space you may need to make changes (to the layout or the technology) depending on the current environment.
For example, you might want the lights in your garage or entryway to come on automatically. One obvious and easy solution would be to replace the light switch with a sensor switch. By thinking about how the space is used and how the switch might integrate with your home automation system you can decide if a sensor switch is the best solution. Consider things like: is the switch in a good location to sense when the space is occupied? Is it likely the switch will be blocked by items stored in the space? Will the switch work with the lights? Will this connect with other home automation systems you have or plan to add?
Things to consider about where home automation technology is installed in shared spaces:
- If the technology uses occupancy sensors ensure they can be tuned to work with your space. They should be able to cover an entire space so no one is missed or narrowed to limit sensors in areas you do not want covered. For example you may not want pets or children to trigger events.
- Wherever possible select systems and technology that will accept input from multiple sensors or are designed specifically to recognize multiple occupants
- Flexibility can be an asset. It might look nicer to build something into the space like replacing a wall socket or light switch with one with integrated home automation features; however, with multiple occupants the way the space is utilized may change more frequently and having the ability to easily move or simply remove the devices from a space quickly can be useful.
- Flexibility can be a problem. Wait, what? Didn’t you just say it is an asset? Well every coin has a flip side and in shared spaces things that change often or are not bolted down can be confusing or can be ‘lost’. Know your space and think carefully about what is more valuable.
To automate or not to automate. That is the question. Waaay back up in the “What” section we talked about how easy it is to get caught up and begin automating everything. Assuming your compulsion is under control there is still a time and a place for everything. Controlling when automation runs and when it doesn’t can be the difference between a good system and a great one.
Deciding when an action occurs is determined by what triggers you set. The more information the system has the more complex those triggers can be and the more nuanced the response can be made. This is one of the tricky balancing acts where more is not necessarily better.
Let’s look at the example of having the lights come on in the garage again. If we change the switch to a remotely controllable one and add two or three motion sensors we can improve the likelihood that the lights will come on when we need them. We may find that the lights are coming on even when the garage door is open and lots of sunlight is flooding in. That is a waste of energy and frankly a bit annoying; so let’s change the motion sensors to multi-sensors that can detect motion and the brightness of the room. We then set the trigger to activate the lights if the sensors detect motion and less than fifty lux of light. For some people that would seem like a reasonable, well thought out solution and they would stop there. Others are scratching their heads right now and saying just put the light switch back. Finally others are thinking: that is a great start but we can also tie in my calendar and adjust the sensors based on when I should and should not be home and the system can send me an alert and those multi-sensors can probably sense temperature as well, let’s set it up so they flash the lights if the temperature drops below freezing in the garage and… you get the picture. For every person there will be a different tipping point for when a trigger is tripped and what action is taken. In a shared environment you should design the system to the middle. The middle is the point where compromise and agreement can be reached so that everyone can be happy with the automation.
Here are our recommendations for when automation should be used in a shared space:
- Just because a system can do something doesn’t mean it should. Watch for the point of diminishing returns
- Monitor the space with the right sensors so automated actions happen in appropriate circumstances
For some people asking why they automate is like asking why climb a mountain – because it is there. In a shared space where not everyone may appreciate or want the automation you need to consider what extra utility, comfort or safety does the automation offer and do the benefits outweigh the drawbacks.
There are many reasons to add automation and they usually fall into one of three categories: comfort/convenience, cost savings and safety/security. The complexity in a shared space is understanding the value for everyone in relation to the costs. Planning the system carefully and understanding the needs and impacts for everyone is the best way to ensure you deliver a good system for everyone.
Our recommendations for considering why when automating in a shared space are:
- Plan carefully. Understand all users needs and how the system may impact their use of the space.
The Advanced Class
In this section we look at some of the more complicated things you can do with your home automation system to make it work in a shared space. While we will try to stick to the principles of ‘simple is better’ these sections by definition tend to be nuanced, layered and therefore more complex.
If you aren’t interested in getting into this level of complexity you can jump right down to the wrap up where we summarize our thoughts and give you some ideas for further reading. Jump down to the Wrap Up.
The Advanced Class covers using artificial intelligence in multi-user environments, meshing systems together with hubs and software, physical control systems and providing a unified interface, and using multiple sensor inputs and facial/biometric/other recognition for custom interactions.
When we talk about artificial intelligence (AI) related to a shared space automation system unfortunately we aren’t talking about a J.A.R.V.I.S like ultimate and aware presence that will know who you are and customize your environment based on that knowledge (wouldn’t that be cool and I am sure it isn’t that far off) we are talking about specialized systems that interpret complex commands. There are technologies that can recognize individuals and complete specific tasks. We will look at recognition systems in the last section “Sensing Who is in the Room” but that is for later.
Artificial intelligence in home automation is a hot area right now. Major players include Amazon, Google, Apple and Microsoft. We may see something from Facebook soon and there are dozens of other smaller players that have interesting and niche products. The main advantage that AI brings is untangling complex commands or situations. The best AI today specializes in translating what you have said into what you probably want by taking into account the context then making ‘smart’ decision or at least correctly interpreting what it is that we actually want. In a shared space having an AI device actually adds a substantial layer of complexity.
Most of the big players don’t even attempt to recognize multiple users at this point and if they do work in a shared environment it is by specifically telling the AI to switch to a different user.
Some key things to remember when using AI systems in a shared environment include:
- These systems can be a serious security vulnerability. Most of these systems can be easily controlled with just voice commands. Keep in mind that any access you have setup (whether it be to your bank or your favourite pizza joint) can be exploited by anyone who knows how the system works. Wherever possible use a pin or passcode or customize the command sequence so that anyone with access cannot abuse that power.
- While they are pretty smart and good at understanding you don’t underestimate the potential risks of a misinterpreted command. You may ask “Order me a pizza” and the pizza guy shows up with your favourite ‘za a short time later or a week from now a goofball bronze pizza statue shows up in the mail.
- With a shared space and no need to actually see the things you are interacting with you can have lots of unexpected consequences imagine you issue the command: “turn the sprinklers on in 3 hours” result: your neighbour/dog/kids/self are soaking wet) or something like “start the laundry in an hour” and who knows what someone loads into the machine in the meantime.
Controller Hubs & Software
You can add a lot of automation to a shared space without a hub. That said, adding a controller hub opens a lot of options and allows you a level of control and flexibility that isn’t otherwise possible.
For example, five years after launching their segment leading product one of the biggest players in smart lighting, Philips, doesn’t geofence for multiple users natively. The control app is a great piece of software and a significant selling feature but it breaks down quickly in a shared space.
Having a great product that works well requires hard work. Having a great product that works well and tracks two people isn’t just harder or even twice as hard it is an order of magnitude more complex; which is why most vendors don’t even try.
Using a controller hub you no longer need (get?) to use the native application. Controllers can be programed with more complicated triggers and actions; especially the controllers that are geared more toward the hacker or advanced users like the Vera or HomeSeer or elite systems like Crestron or Savant. These hubs provide a relatively simple web interface but also allow you to dig deeper into the guts of how the system works and add plug-ins or even write code to really customize. Even the most basic hubs usually have more functionality baked in than even the most complex native software because by their nature they are designed to work across vendors and/or with a wide range of devices. Working with a controller you can layer triggers like: when person one and person two are away and time of day is night and noise level is less than 50 decibels at sensor one and three… You can also layer the actions that occur like: turn off bedroom lights and basement lights and set the system to away. You get the picture.
The challenges with hubs are that they are complex and can be complicated to configure, more prone to malfunction and the more devices you connect the more difficult they can become to manage.
Somewhere between using the native application and a controller hub are a breed of software applications that can chain together conditional statements and trigger outcomes. While not as powerful as a controller hub they can do a lot and sometimes are significantly easier and cheaper (IFTTT one of the best solutions is free!) to setup. To replicate the case of turning off your lights when everyone is out with software service like IFTTT you would need to do something like: If my thermostat senses that no one is home then turn off all of my hue lights. So, you can merge together sensors and algorithms from one vendor to carry out actions for devices from another vendor.
The major challenge with software services is that they are limited to what devices and services they connect with and what actions and triggers they can access for those services. They also require internet connectivity to work.
Our recommendations for control hubs and services in a shared environment are:
- Carefully understand your needs before you select a hub and the devices/services you will control with them. Keep to the principles outlined above and keep the system as simple as possible while keeping your options open to grow and adapt
- Start slow with something simple. Make sure your first devices and rules are working well and reliably before moving on to more complex projects
- Do your research and try to predict your future requirements carefully. Know what the strengths and weaknesses of controllers and services are before you lock yourself into an environment or ecosystem.
Things can get out of control quickly if you have a mashup of different devices and services all managed by their own applications. The scenario becomes a special kind of crazy when you add multiple people controlling multiple systems simultaneously. One of the best ways you can bring order to the chaos is by unifying the interface and eliminating multiple contact points as much as possible.
Using a control hub with a dedicated control application can be an easy way to achieve a unified interface. The interfaces can be via a combination of web and/or apps. It is becoming increasingly popular to only offer control via apps with no web or other interfaces. It is always a very good idea to make sure you can control the hub you select the way you want. Think about the big picture so that you are not left hanging with a system that can only be controlled by one phone in the family or where you need to go out and buy a bunch of new gear.
Controllers that can only be managed via an application are becoming increasingly popular because consumers are looking for easy to use systems that work with their phones and tablets. These systems often have beautiful graphics and slick interfaces. Keep in mind that simplicity often means sacrificing functionality. However, as long as an app only hub controls everything you need, they are particularly well suited for use in shared spaces because of their ease of use and relatively fewer functions or options for modifying the configuration and layout.
The weak point for control hubs, especially the more sophisticated ones is that their control applications and web interfaces tend to be more complicated and often glitchy as well. Many vendors in the more advanced amature hacker/maker segment tend to have a lot of open source development including extensive plugins and scripts that do amazing and complicated things and unfortunately are not always the most stable or reliable solutions. On the plus side they are typically supported by great communities who are quick to fix problems that impact lots of users or give you the tools to go in and fix things yourself if you are inclined.
One final way you can take your system to the next level and really improve and simplify automation in a shared space is with a good ol’ fashioned physical controller. These range from a basic candy bar type remote control similar to what you find with any other electronics to pretty much any crazy configuration you can imagine. It is common for elite systems to have a dedicated portable tablet or tablets built into key areas with customizable controls. Custom control interfaces can also be built for multi-purpose tablets but keep in mind that in a shared space if you are using a physical remote control it is probably because at least some of the people using the space are looking for something simple and consistent. A tablet that walks away or is being used for something else will probably result in as much or more frustration as an app on a phone. One way higher end systems tackle this issue is using an existing tool, often a television but sometimes a multi-function remote or other device, that everyone is comfortable with and layering in a nice control interface.
Our recommendations for unified interfaces in a shared space are:
- Use a control hub wherever you can to provide a central point of control
- Ensure everyone is accessing the same controls and make the interface work for the lowest common denominator
- Consider using a physical controller. In a shared space where there are those who don’t like to use technology go with something that actually has physical buttons works well.
- Use a remote that can be programmed with macros and make sure those macros can reset themselves so if pieces of a scene get out of sequence people without technical know-how can simply press the button to turn everything off and back on again and it should all work.
- Make sure everyone who is using the system understands what it is doing or at least how it should work. Pay particular attention to any devices that might have long delays. Long delays for people who expect things to happen right away can be as little as a couple of seconds! Make sure they know that for example: the receiver will take a count of ten before it is ready to go and you will see a picture on the tv rather than how it used to jump right up.
Sensing Who is in the Room
Determining who is in the room seems like an obvious first step for automating a shared space. Unfortunately it is far harder to get working than it sounds. There are systems that use biometrics like facial recognition to identify you like the Microsoft Kinect or NetAtmo Welcome. Another solution might be using tokens or a personal device to recognize a specific person (or at least there belongings) are in a space. Probably the most basic solution is to use something you know – like a passcode to identify who you are.
Knowing who is in a space allows you to customize that space. For example using a keypad door lock that when unlocked with a personal code opens the door and also lets the system know who has entered the space. The system can then start a series of personalized events such as resuming music that was playing using specific speakers, adjusting lighting to lead to that individual’s room or turning on a device like the television or coffee maker.
Systems are setup to recognize people everyday but our recommendation in a shared space is actually that you avoid it unless there is a real driver for it. If you do implement personalized systems we recommend you keep them as simple as possible. Either use the systems for the specific task they were designed for like with the Kinect or Welcome or keep the customization non-invasive such as turning on a light in a personal space when you arrive home. By keeping things simple you reduce the requirement for complex if/then analysis and the system will need to account for fewer overlapping user requirements.
Our recommendations for sensing who is in the room in a shared space:
- Don’t do it unless there is a real driver for customized automation
- Use something you know (like a passcode) rather than something you are (biometrics) or something you have (a phone or token). This allows you to consciously decide when you want a customized action to occur.
The Wrap Up
Well congratulations! At this point you have covered a lot of ground and have a pretty good understanding of automating a shared space including the complexities and pitfalls. You also have some good ideas from our recommendations about how to avoid or work around those challenges.
One of the biggest challenges we haven’t discussed yet is the simple fact that there are no perfect solutions or right and wrong answers. There are an infinite number of challenges you might face and an equally vast number of possible ways you can work through them.
The key things we think you should consider when designing an automation system for a shared space include:
- Keep it simple
- Plan what are are doing well
- Keep everyone who uses the shared space informed and seek their feedback
- Test what you can before you put it in place
- Build slowly – one new device/system at a time if you can
Wait, there is nothing technical in those recommendations! Well, that’s because when you are building a system for a shared space the people are the most important components. It is easy to get lost if you focus on the technical components.
If you want something a little geekier here are some of the key kinda-technical considerations:
- Keep it simple – we can’t get enough of this one. In this case we are specifically referring to the concepts of: use one platform as much as possible, keep your triggers and actions as close as possible and eliminate variables and conditions as much as you can
- Try to use multi-purpose sensors/devices/services wherever you can so you get the best bang for your buck, smooth integration, and reduced complexity
- If you have more than a couple devices/platforms seriously consider using a hub to coordinate services, provide additional functionality and help with a unified control interface
We hope you have found this interesting and helpful. If you are interested in more of the same or have questions about some of the technology we reference here check out our other articles. A good place to start is the “Start Here” section if you haven’t been there already – check it out! Alternatively there is the ‘Everything Else” section if you are looking for definitions for anything you didn’t understand. The “Stuff We Love” area is a good place to start if you are inspired and want to jump in and give some of these ideas a try and are looking for great gear.