EMG Armband

Developing an EMG system that can understand human hand motions.

What?

Our project focuses on reading Electromyography (EMG) data using sensors. We follow these steps:

1. Amplify the EMG signals
2. Filter the data
3. Perform ADC conversion
4. Conduct frequency analysis
5. Feed the analyzed data into a ML model to recognize and classify hand motions

Project Structure

Our codebase is organized as follows:

• /ml: Contains the machine learning model code
• /muskel: Houses the ESP32 package code for reading data, performing windowing, and conducting frequency analysis
• /EMG-Amplifier: Contains files for our custom EMG amplifier, built from scratch

Setup Instructions

Setting up the muskel

cd muskel
idf.py build
idf.py flash monitor >> dataset_file.csv

Machine Learning

cd ml
python3 model_cnn_lstm.py  # Starts the training run
python3 inference.py       # Starts the inference

Additional Resources

• Project Report
• Blog Post: How Does EMG Armband Work?

Thanks to upsidedownlabs for sponsoring and supporting our research!

---

Made with ❤️ by the members of SRA