On-Demand

Boost your Edge AI development with Zephyr and TFLite Micro

112 views
November 01, 2021
The nRF52 Series is supported in the Renode framework. Renode is an open source simulation framework that allows you to run, test and debug embedded software in various scenarios: from simple single-node tests to complex wired and wireless connected systems.

Adopted by Continuous Integration (CI) of many popular projects, like the Zephyr RTOS or Google's TensorFlow Lite for Microcontrollers (TFLite Micro), Renode helps developers ensure their software correctness at every commit, without the need to access the real hardware.

In the webinar, we will present how to quick start your development with Renode, from the very basic debugging to complex machine learning flows analyzing data from sensors. We will also show some recent developments in the nRF52840-based Arduino Nano 33 support, implemented in collaboration with Google and the TFLite Micro team.

Agenda:
- Intro to Renode, Zephyr and machine learning - how they are related
- Basic usage - basic single-machine setup, loading a Zephyr binary, interacting with Renode
- Interactive debugging with GDB
- Feeding data to sensors
- Automated testing in Renode
- Arduino Nano 33 visualisation within Renode
- TFLite Micro CI setup
- Interactive Google Colab notebook for AI workflow testing on the nRF52840 SoC

This is a recording of a live webinar held on October 28, 2021. The Q&A is a concatenation of both the morning and evening sessions. The presentation can be downloaded here (direct download).