For many engineers, the artificial intelligence of things (AIoT) is a gateway to introducing a diverse range of new features into their products, using on-device AI.
In our Edge of Now report, engineers shared the most important features that AI would enable for them. Enhanced performance was highest on their list, followed sequentially by ease of use, security, flexibility, and improved user interfaces.
But how do these things manifest themselves? As our previous blog discussed, the AIoT incorporates so many different combinations of technology and features for each use case – so how do we know which capabilities will be deployed most widely?
This blog is an attempt to clarify that. We look at six of the most exciting features that the AIoT can offer in electronics of all kinds, and outline which will help to drive new and innovative use cases.
A deep dive into the features
1. Sensor processing
To make sense of the data that they receive from sensors, most electronics rely on cloud-based transmission of data out of the home and to a third party. On-device processing, by contrast, allows for data to be interpreted without any external broadcast.
In freeing themselves from the shackles of cloud-based dependency in their products, engineers can expect improvements in latency, speed and privacy – as data can be acted on as soon as it is received.
2. Remote monitoring
Conventional electronics with sensors are always ‘listening’, monitoring their environment for the stimuli to a set task using their full power supply, – again depending on transmissions to and from the cloud.
AIoT devices, by contrast, can observe their environment even while in a low-power state, awaiting triggers that will activate other embedded components. This is inherently more energy efficient as full capabilities are only used when required.
3. Data integration
The aforementioned capacity for higher performance, and the ability to monitor environments, enables AIoT appliances to combine data streams. The built-in intelligence allows devices to integrate multiple data streams, contextualising the meaning of specific stimuli within any given environment. This enables more sophisticated decisions and responses to be managed locally.
For example: if you can teach a smart speaker to recognise what a human looks like when they stumble or fall over, and the sound of sudden impact or a call for help, it can combine these events to ‘understand’ that it should call emergency services.
4. User Interface improvements
Most ‘smart’ devices currently rely on a touch-based interface. While this has been deemed a simple, streamlined solution, some tasks, such as setting alarms on your phone, are still complex multi-step processes.
Voice technology addresses this issue, and on-board processing will enable new heights in the recognition and execution of spoken commands. Not only does this mean that users can execute tasks without reliance on touch, but it also allows for a more efficient user experience. What’s more, it also addresses the very real concerns shared by many about the hygiene risks associated with touch screens that have surfaced since the Covid-19 pandemic.
5. Communications
None of the aforementioned features are taking place in isolation. AIoT-enabled devices can share their data with one another, taking advantage of their collective intelligence to execute complex tasks as a group.
For example, if someone were to say, “Hey Google, goodnight”, the smart products in the home would communicate through the Wi-Fi connection, directly to your smart switches, which would then turn the lights off. Gadgets are no longer mutually exclusive; the smart home network is a prime example of technologies combining to collectively control an entire environment.
6. Predictive maintenance
We’ll also see longer product life, too. The data from AIoT devices will be able to show when machines are underperforming, malfunctioning, or reaching the end of their lifespan, enabling the user to intervene. In some environments – manufacturing, for example – this is significant for preventing injury, maintaining peak performance, and avoiding threats to day-to-day operations.
Looking to the future
With all these new features, the AIoT is offering engineers greater scope to build in features that will significantly improve convenience, safety, intelligence and interactivity for people across a range of sectors. This is thanks to the high functionality, small scale, and adaptability of the fundamental electronics underpinning these designs.
Read this year’s Edge of Now report here to find out more about how the possibility of AIoT learning is becoming more and more of a reality.
