WiGuard - Model

Wireless sensing based health monitoring system.

How to run

If you use a virtual environment, name it .venv or .conda as it is already added to the .gitignore file.

Then, create a .env file in the root directory of the project and add the following content:

MQTT_BROKER_HOST=localhost
MQTT_BROKER_PORT=1883

Install the dependencies and run the application:

pip install -r requirements.txt
python main.py

To run the project on a GPU, you need to install the correct version of PyTorch. For example, to install PyTorch 2.5.0 for CUDA 12.4, run the following command:

pip3 install torch==2.5.0 --index-url https://download.pytorch.org/whl/cu124

Workflow

graph LR

ESP32-->NanoMQ
Intel5300-->NanoMQ
NanoMQ-->Cloud
Model[⭐Model]-->Cloud
Cloud-->Web
Cloud-->Client

Loading

Utils

Visualize csi data:

python -m wiguard.live csidata.csv

Predict one csi file:

python -m wiguard.predict csidata.csv

Subscribe to the MQTT broker and show:

python -m wiguard.mqtt

APIs