AI on STM32: fan anomaly detection and classification on current sensing with NanoEdge AI Studio

In this demo we use NanoEdge AI to implement edge AI anomaly detection and classification capabilities in a fan device. Discover NanoEdge AI Studio at:

For the experiment we use a Cortex-M4 STM32 microcontroller. Anomaly detection learning and inference are done locally in the microcontroller without any pre-trained model required. In addition a classification library is pre-trained with NanoEdge AI Studio.

Be sure to take a quick tour of the brand new NanoEdge AI Studio V3 and learn how it can help simplify the creation of machine learning capabilities at the edge with our popular STM32 family of microcontrollers. Try it today at:

To find out more information about AI solutions at STMicroelectronics:

#edgeai #tinyml

0:00 Intro
0:21 Hardware setup
0:40 Demo – Learning phase
0:55 Demo – Training the anomaly detection library
1:10 Demo – Nominal detection
1:19 Demo – Anomaly detection (clogging)
1:38 Demo – Anomaly detection (friction)
1:49 Recap
2:00 Outtro

SUBSCRIBE to our YouTube channel for more content like this:

FOLLOW STMicroelectronics on Social Media:


%d bloggers like this: