HAVOC-Server

Source code on GitHub.

README

HAVOC-Server: Temi Monitoring System

This project is a monitoring system for the Healthcare Assistant with Video, Olfaction, and Conversation (HAVOC) program running on the Temi robot. It uses computer vision and machine learning to detect falls, classify smells, and generate daily reports. It provides real-time metrics and visualizations through a web interface, making it ideal for applications in healthcare, safety monitoring, or robotics research.

!image

Features

• Fall Detection: Uses YOLO (You Only Look Once) for real-time detection of falls with multiple methods (bounding box, pose keypoints, and bottom fraction).
• Smell Classification: Employs a K-Nearest Neighbors (KNN) model to classify smells based on sensor data.
• Real-Time Streaming: Streams live video and sensor data from the Temi robot via WebRTC.
• Daily Reports: Generates and sends daily reports with metrics and t-SNE visualizations of smell data.
• Web Interface: Provides an interactive dashboard for monitoring, recording, and visualizing data.

Installation

• Clone the repository:

``bash git clone https://github.com/innovationcore/HAVOC-Server.git cd HAVOC-Server `

• Create a virtual environment:

`bash python -m venv venv `

• Activate the virtual environment:

- On Windows:

`bash venvScriptsactivate `

- On macOS/Linux:

`bash source venv/bin/activate `

• Install the required dependencies:

`bash pip install -r requirements.txt `

• Additional Setup:

- Download the YOLO model weights (yolo11n-pose.pt) and place them in the yolo_weights/ directory.

- Set up environment variables in a .env file:

`plaintext TEMI_EMAIL_SENDER=your-email@gmail.com TEMI_EMAIL_PASSWORD=your-app-specific-password TEMI_EMAIL_RECIPIENTS=recipient1@example.com,recipient2@example.com TEMI_REPORT_TIME=20:00 `

- Ensure the static/newSensor_training.csv file exists for smell classification training data.

Running the Application

To start the server, run:

`bash python server.py `

• The Flask server will start on http://0.0.0.0:8133, serving the web inter...