Pete Warden Points to TinyML as a Quick, Cheap Alternative to Smart Meter Deployment

Pete Warden has highlighted the potential for TinyML and screen-scraping technologies to transform the world of "dumb" appliances into smart systems, reading everything from numeric displays to analog dials automatically.

The move to smart meters for gas, electric, and other utilities is ongoing, but troubled: Demand for smart meters outstrips supply, fitting them requires a disruption in utility provision, and there are concerns about their reliability and security. It'd be a lot easier if it were possible to simply and non-invasively inject some smarts into existing hardware — and that's what engineer Pete Warden, best known for books The Public Data Handbook and The Big Data Glossary, predicts will happen.

"This image shows a traditional water meter that’s been converted into a web API, using a cheap ESP32 camera and machine learning to understand the dials and numbers," Warden writes, referring to a screenshot from the open-source AI-on-the-Edge-Device project. "I expect there are going to be billions of devices like this deployed over the next decade, not only for water meters but for any older device that has a dial, counter, or display."

"Pointing a small, battery-powered camera instead offers a lot of advantages. Since there’s an air gap between the camera and the dial it’s monitoring, it’s guaranteed to not affect the rest of the system, and it’s easy to deploy as an experiment, iterating to improve it."

Warden points to the release of TensorFlow Lite Micro as a tipping-point, allowing machine learning to be deployed quickly and easily on extremely low-cost and low-power microcontroller devices — TinyML. "What I’d love to see, Warden adds, "is some middleware that understands common displays types like dials, physical or LED digits, or status lights. Then someone with a device they want to monitor could build it out of those building blocks, rather than having to train an entirely new model from scratch."

Warden's full article is available on his website, while AI-on-the-Edge-Device can be found on GitHub under the permissive MIT license.