Motion classification application on an Arduino board supporting multiple IoT protocols

This is a motion classification application for an arduino board (using embedded ML capabilities of the hardware) that sends events to the external world either:

• by sending the events through MQTT
• by sending events with the CoAP protocol
• by declaring a "Web Thing" interface, allowing to monitor it via a WebThings Gateway running on a Raspberry Pi

It was my final project at the Summer School on Enabling Technologies for IIoT by University of Pisa, 2022. Check the full report here.

Requirements

Hardware:

• Arduino Nano RP2040 connect
• Raspberry Pi for running the WebThings Gateway software (Optional)

Libraries:

• WiFiNINA
• STM32duino X-NUCLEO-IKS01A3
• ArduinoMqttClient
• CoAP-simple-library
• webthing
• ArduinoMDNS

Example setup:

Features

• Send the current motion status through an MQTT topic to a broker running on a cloud server. (See source code)

  

• Send the current motion status as a CoAP PUT request to a CoAP test server running on a cloud server. (See source code)

  

• Send the current motion status to a WebThings gateway running on a Raspberry Pi on the local network. (See source code)

  

References:

• Using the IMU Machine Learning Core Features of the Nano RP2040 Connect
• Article about Coap by Adafruit
• WebThings Framework
• Shawn Hymel's tutorial on WebThings Framework
• My starred repositories for this project