The concept of a smart home environment has been around for a while, with Google Homes and Amazon Alexas spearheading mainstream acceptance of the idea. These technologies mark a progression in modern domestic living – but do they really fulfil consumer expectations of genuinely smart homes.
Such devices are typically standalone, collect limited data, and falter without an internet connection – not to mention their reliance upon the cloud. Without help, the futuristic residences we see in films will remain confined to science fiction.
A word to the wise
Much of this is due to the limitation of sensors. While voice control represents technical progress, in order to realise a truly intelligent home, we can elevate voice by combining it with other forms of data capture, all working together to deliver seamless home automation.
This can be achieved by using sensor arrays that provide devices with the breadth of data required for them to better contextualise user behaviour. For truly intelligent homes, we’ll need architectures that’ll leverage this rich data capture for greater personalisation.
Bringing the home to its senses
Our homes could be equipped with a diverse array of sensors within each application. Image data alone provides vast amounts of additional information, adding another dimension to a device’s ability to learn and understand. Combine this with acoustic sensors, gas sensing and 3-D mapping, and you’ve got futuristic architectures that’ll seamlessly adapt based on user behaviour, responding to our physical spaces and the actions we take within them.
Using this data, devices will become more intelligent, recognising how to tailor their environment by sensing who’s present, and changing the light, sounds, temperature, and safety features of the home accordingly.
In other words, our devices won’t just observe the environment. They’ll change it for the better — with or without direct input.
A two-pipe problem
This approach holds significant potential, but there’ll be a need to address two key technical and ethical challenges: user privacy and sustainability concerns.
There has always been a concern surrounding the privacy of smart speakers: 91% of US consumers say they’re concerned about external parties listening in. Recent examples include Amazon’s Astro robot being accused of “full-scale data harvesting,” and Ireland’s Data Privacy Commission sounding the alarm over Facebook’s smart glasses.
Enhancing the sensor array of intelligent devices may exacerbate those concerns: the devices, almost by definition, want more data and will be more effective at capturing it. The onus is on brands to set consumer minds at ease by balancing this approach with robust privacy by design.
Then there’s energy consumption, and its implications for the environment. COP26 has reinforced the need to make devices more efficient, so we’ll need to ensure future architectures operate sustainably while still offering powerful continuous processing.
Giving an edge to smart homes
We can resolve these issues by embedding AI into the devices themselves, contributing to the Artificial Intelligence of Things (AIoT). By utilising edge processing, users can keep their data local while also reducing energy consumption to solve the two-pipe problem.
Processing data on-device means minimising external communications beyond the home environment, drastically reducing privacy risks on the cloud. Embedded intelligence will mean that devices can adjust automatically, powering on and off as appropriate – resulting in both a slashed electricity bill and decreased strain on the environment.
In this manner, we can start to bring truly smart, futuristic homes into the realm of science fact. Ambient sensing, facilitated by a comprehensive suite of sensors that can paint a picture of their environment, will enable smart devices to start delivering a level of user experience that was once a pipe dream.