Analog Devices Brings tinyML Machine Vision to Embedded World 2022

One of the best parts of Embedded World 2022 was having the opportunity to stop by the stands of big names we know, and see their products in action. Analog Devices had a number of different demonstrations on display, but the two that really caught our eye were based on Machine Vision and retroactive smart solutions.

We began by speaking with Ole about a very compact little edgeAI tinyML system for facial detection.

Central to this project is the MAX78000 series MCU, which is available in a number of different development form factors and boards. While the models are trained elsewhere, the on-device inference was quick, and it's impressive seeing facial recognition on such underpowered hardware!

We also spoke with Andrew about a mod turning a regular fitness bike into a smart fitness bike, via two strain gauges attached to the crank arms feeding data to a MAX11108 SAR ADC and a MAX32666 MCU communicating via Bluetooth Low Energy to a smartphone app.

There's no specific write-up of this project, though it is mentioned in Analog Devices Sensor Expo handbook, however, this isn't really the point. Seeing disparate parts put together like this to create working demos opens up your mind to what you might make with them. More information on all of the technology mentioned in this blog post are available via their individual links, with further information available on the Analog Devices website.