The advent of ‘ambient sensing’ in the smart home by Mark Lippett, CEO, XMOS
For the everyday consumer, when you talk about the smart home, the image of a suburban HAL 9000 still lingers — speak to any appliance and it’ll do your bidding.
The reality in the present day is that smart homes just aren’t… well, smart. Deloitte cites four main categories of connected device in the everyday home: computers, entertainment, health, and home utility. Useful, familiar, and interconnected — but intelligent?
Well, not exactly. For starters, these devices have been around a long time. The first iPad launched in 2010; Philips first-gen Hue lightbulbs went on sale in 2012, with the first Echo following a year later. As convenient as these devices can be, there is a risk of mistaking ease of use and novel interaction with genuine intelligence.
What’s more, nowadays consumer smart home products don’t extend too far beyond Alexa or Google Home — and if they do, it’s within an enclosed ‘family’ of products. Manufacturers are keen to maximise their streams of customer data without sharing it with their competitors, and that means absorbing you into a system of app and hardware usage that isolates you from some devices while encouraging the use of others.
Whilst such devices offer a streamlined, familiar user experience, honed by years of consumer use, they don’t paint a conjoined, truly intelligent user experience within the home.
Thinking about how our devices understand the home is key to changing that. It’s not necessarily about forcing Google and Amazon devices, for example, to work together — it’s about combining different sensors to capture a more meaningful dataset for smart devices to work with.
Firstly, that means thinking about the results that you want to achieve. Take the humble light bulb, for example. We understand that a microphone in a light bulb allows you to turn it off through a smart home speaker. But if it had a light sensor that could detect human presence, ambient light, and adjust the bulb’s output accordingly, would we need to?
We also know that our smart home assistants recognise voices. But do we need a more sophisticated voice system that can recognise who is speaking and tailor the response to that voice accordingly? The same goes for CCTV. Sure, they can detect movement, but what if they could capture specific visual events, like the fall of an elderly relative, or an unknown person present?
These are the sorts of questions that need answering for the smart home to take intelligence to the next level.
Barriers to the future
Of course, intelligence requires a level of complexity. The more sophisticated that smart home devices become, the more important it is to incorporate artificial intelligence (AI) that can comprehend the different types and growing volume of data that such a device can capture and turn it into action.
Such complex technology does not come without drawbacks. AI functionality is founded on data — the more intelligent the device, the more data it requires. Sophisticated, futuristic smart home devices would normally be required to collect and share large amounts of information with a central “brain” to process it, like the cloud, which understandably raises privacy concerns among consumers. Individuals already raise concerns about the possibility of smart speakers listening in to conversations and no one would want more of their personal information being shared with external servers that are beyond their control.
Latency issues also pose another barrier to the seamless smart home of the future. We have all experienced the pure frustration created by lag or connectivity issues. On slower connections, smart speakers can already take a considerable amount of time to receive requests, process them in the cloud and execute the desired action. As the number and complexity of smart devices continues to increase, these issues will continue to grow. The cloud simply won't be able to scale with the sheer amount of data produced and irritating delays will continue to haunt the smart home.
If we are to introduce intelligent ambient sensing into the smart home a change of tack is needed. Step forward the artificial intelligence of things (AIoT).
An intelligent home
AIoT technology brings together AI and IoT — creating hardware capable of collecting and processing data locally. Cutting out the need for cloud connectivity, AIoT minimises the security and latency issues previously associated with smart technology. Collected information is processed on the smart devices themselves – keeping personal data secure and executing actions almost immediately.
This technology could have huge impacts on the smart home, allowing devices to work together instantaneously, creating a truly intelligent home. Imaging, acoustics, 3D mapping and even gas sensitivity could all be integrated into smart devices, allowing all variables to be monitored and automatically acted upon. Homeowners would be able to have a whole new level of control over their homes, far extending current voice-based offerings.
While current smart technology was impressive in its infancy, we are reaching the point where the next step is needed. The novelty of voice assistants is waring off and frustrations with current offerings are building. The introduction of ambient sensing, in combination with the adoption of AIoT, will revolutionise the sector. As technology improves and demand increases, the advent of truly futuristic, intelligent homes may be a lot sooner than we think.
For further information please visit https://www.xmos.ai/