TinyNML

Welcome to TinyNML, a growing collection of embedded machine learning projects developed in the Neuromechatronics Lab at Carnegie Mellon University. This repository focuses on deploying ML models on resource-constrained microcontrollers like the Raspberry Pi Pico and other edge devices.

Whether it's motion recognition, gesture control, biosignal processing, or edge AI inference pipelines, this repository aims to provide a collaborative platform for lab members to prototype, share, and deploy their models on embedded hardware.

🧠 Project Goals

• Build scalable and modular TinyML applications on embedded hardware
• Combine mechatronics, biosignals, and ML inference for real-world applications
• Provide a collaborative platform for lab members to prototype and share models

---

🚀 Current Projects

Project Name	Description	Language	Hardware
pico_motion_classifier	Real-time circular motion classification using MPU6050	C++	Raspberry Pi Pico W + MPU6050 + SSD1306
gesture_oled_demo (coming soon)	Accelerometer gesture display using Edge Impulse on CircuitPython	CircuitPython	Pico W + MPU6050 + OLED
keyword_detect (coming soon)	Audio keyword spotting	CircuitPython	Pico + MEMS microphone

---

⚙️ Getting Started with Pico SDK

---

These steps guide you through setting up the toolchain. If using a Windows PC, it is recommended to use Ubuntu under WSL2 for compiling UF2 firmware for the Raspberry Pi Pico.

🔧 1. Install Dependencies

Run the following in your terminal for compilation tools:

sudo apt update
sudo apt install -y cmake gcc-arm-none-eabi libnewlib-arm-none-eabi build-essential git python3-pip

Clone the Pico SDK and TinyNML repositories:

git clone https://github.com/raspberrypi/pico-sdk.git
git clone https://github.com/Neuro-Mechatronics-Interfaces/TinyNML.git

(Optional) Set the PICO_SDK_PATH environment variable:

• Windows:

  export PICO_SDK_PATH=/path/to/pico-sdk

• Linux:

  echo "export PICO_SDK_PATH=~/pico/pico-sdk" >> ~/.bashrc
  source ~/.bashrc

⚙️ Getting Started with CircuitPython

---

1. Download the latest CircuitPython firmware for your Pico W from CircuitPython Downloads.
2. Press and hold the BOOTSEL button on the Pico W while connecting it to your computer via USB. This will mount the Pico as a mass storage device.
3. Copy the downloaded .uf2 file to the Pico W. It will reboot and appear as a new drive named CIRCUITPY.
4. Install any required libraries:
   • Download the latest Adafruit CircuitPython bundle.
   • Copy and include only the necessary libraries from the bundle to the to the lib folder in the CIRCUITPY drive.