Efficient sensor data processing with MicroPython (presentation from PyCon ZA 2024)

[jonnor]

Hi everyone.
If you have tried to process a lot of data with Python on a microcontroller, you may have noticed that is not the fastest. If trying to go beyond a few 100 Hz data rate, you may hit challenges. But MicroPython actually has a bunch of great tools that make it possible to greatly improve efficiency, while still keeping a clean high-level approach to your program. In the presentation below I go through some of these, such as:

• Native and Viper code emitters
• C modules

And discuss briefly applications to processing accelerometer data, audio data, etc. This is based on my experience from the emlearn-micropython project, a Machine Learning and Digital Signal Processing library for MicroPython.

Sensor data processing on microcontrollers with MicroPython (PyCon ZA 2024)

Video recording
https://www.youtube.com/watch?v=vBiji9IFJJs

Slides and materials
https://github.com/jonnor/embeddedml/tree/master/presentations/PyConZA2024

Happy to take any comments and/or questions :)

[scruss]

lovely work!

your emlearn link seems a little broken, tho

[jonnor]

Thanks! Link fixed now. I had forgotten https:// so it became a relative link based on this page URL...