Machine Learning on Microcontrollers With Zephyr and Emlearn

Explore machine learning deployment on microcontrollers through this 44-minute conference talk from the Linux Foundation's Open Source Summit. Discover how modern machine learning enables automatic extraction of valuable information from sensor data and learn about the feasibility of deploying ML systems to low-cost embedded devices in the emerging "TinyML" field. Understand the emlearn open-source project, which provides a Python library for converting scikit-learn or Keras models into efficient C code that runs on any device with a C99 compiler. Learn about real-world applications across various sensor modes including audio, vibration, power-line, and radar systems. Examine how Zephyr RTOS, a comprehensive open-source operating system for microcontrollers, supports low-power operation, communication protocols, and standardized sensors APIs to create an attractive platform for TinyML applications. Gain insights into key features and tools provided by the emlearn library and see practical demonstrations of machine learning tasks implementation on microcontroller platforms.