🚨 Emergency Vehicle Detection & Traffic Control System

Real-time emergency vehicle detection using YOLOv5 with automatic traffic light priority control.

📋 What is This?

Proof-of-concept system that detects emergency vehicles (ambulance, fire truck, police) using computer vision + audio detection, then automatically gives them traffic light priority.

Key Technologies:

• YOLOv5 for real-time object detection
• 4-microphone array for directional audio
• Dual verification to reduce false positives
• Raspberry Pi 5 with GPIO control

⚠️ Prototype Disclaimer

This is an academic prototype for educational purposes only.

Suitable for:

• ✅ Academic demonstration
• ✅ Technology proof-of-concept
• ✅ Learning embedded AI systems

NOT suitable for:

• ❌ Production deployment
• ❌ Real traffic scenarios
• ❌ Safety-critical applications

🚀 Quick Start

For Users

See: THONNY_QUICKSTART.md

Quick summary:

cd ~/final-project
source venv/bin/activate
python3 src/deployment/main.py

For Hardware Setup

See: HARDWARE_SETUP.md

Includes GPIO pin mappings, component specs, and wiring details.

For Training Your Own Model

See: notebooks/training_notebook.ipynb

Train on Google Colab with your own dataset.

📂 Repository Structure

final-project/
├── dataset/                 # Training dataset 
├── exp_results/             # Training result
├── images/                  # Demo documentation
├── models/                  # Trained YOLOv5 model
├── notebooks/               # Training notebook
├── src/deployment/          # Main application code
├── .gitignore               # Ignore file
├── HARDWARE_SETUP.md        # Hardware specifications
├── README.md                # This file
├── THONNY_QUICKSTART.md     # User guide
└── requirements.txt         # Dependencies

📚 Documentation

Document	Purpose
THONNY_QUICKSTART.md	Setup and run the system
HARDWARE_SETUP.md	Hardware specs and GPIO mapping
dataset/README_DATASET.md	Dataset information
notebooks/training_notebook.ipynb	Model training guide

🎓 Academic Context

Final Project - Telkom University
Course: Thesis
Supervisors: Yulinda Eliskar & Rita Purnamasari
Year: 2025

Demonstrates:

• Computer vision in traffic management
• Multi-modal sensor fusion
• Real-time embedded systems
• IoT integration

📊 Performance

Metric	Value
Detection Accuracy	~87%
Response Time	~75 milliseconds
System Uptime	~2 hours continuous

🛠️ Technology Stack

Python 3.11 • PyTorch • YOLOv5 • OpenCV • PyAudio • Raspberry Pi 5 • GPIO

📞 Contact

Author: Hady Sadya
Email: hady17306@gmail.com
GitHub: @hadysadya

---

Version: 1.0.0 | July 2025