🚦 AI-Powered Adaptive Traffic Light System

🔥 Overview The AI-Powered Adaptive Traffic Light System is a smart traffic management solution that dynamically adjusts traffic lights based on real-time congestion analysis. By leveraging computer vision, machine learning, and IoT communication, this system enhances urban mobility, reduces congestion, and optimizes traffic flow.

⚡ Key Features

✅ AI-Based Traffic Analysis – Uses a trained YOLOv5 model to detect vehicles and predict congestion patterns. ✅ Dynamic Traffic Light Control – Automatically adjusts signal timings based on live traffic density. ✅ IoT Integration (MQTT) – Uses an MQTT-based communication system to update traffic lights in real-time. ✅ Live Analytics Dashboard – A Flask + AWS dashboard visualizes congestion hotspots and system performance. ✅ Edge Processing on Raspberry Pi – Runs AI inference directly on a Raspberry Pi 5 with the Hailo AI Kit for real-time traffic detection.

📌 System Architecture

1️⃣ Traffic Data Collection

Streams live traffic footage from public cameras (or local webcams). YOLOv5 detects vehicles (cars, buses, trucks) in each frame. 2️⃣ AI-Based Traffic Analysis

Analyzes congestion levels based on vehicle count. A TensorFlow-trained ML model predicts future traffic patterns. 3️⃣ MQTT-Based Traffic Light Control

Traffic light duration adapts dynamically based on congestion. Uses Mosquitto MQTT to communicate with IoT traffic lights. 4️⃣ Live Dashboard & Analytics

A Flask + AWS-hosted web dashboard visualizes traffic conditions. Displays real-time vehicle counts, congestion levels, and traffic trends. 🛠️ Setup & Installation

1️⃣ Clone the Repository git clone https://github.com/kevinomerkilic/smart-traffic-ai.git cd smart-traffic-ai 2️⃣ Set Up Virtual Environment python3 -m venv yolov5-venv source yolov5-venv/bin/activate pip install -r requirements.txt 3️⃣ Download Pretrained Model wget https://github.com/ultralytics/yolov5/releases/download/v7.0/yolov5s.pt 4️⃣ Run YOLOv5 for Traffic Detection python detect.py --weights yolov5s.pt --source "https://your-traffic-camera-url.m3u8" 5️⃣ Set Up MQTT Broker (Mosquitto) Install Mosquitto on the Raspberry Pi:

sudo apt update sudo apt install mosquitto mosquitto-clients Run the broker:

mosquitto -v Test publishing messages:

mosquitto_pub -h localhost -t "traffic/lights" -m "green" 6️⃣ Run Flask Dashboard python app.py Access it in a browser at:

http://localhost:5000 📊 Live Dashboard Preview

The Flask-based dashboard provides: ✔ Real-time congestion updates ✔ Vehicle count graphs ✔ Adaptive traffic signal status ✔ Traffic trend predictions

🤖 How AI Model Works

The system uses a YOLOv5 model trained on a custom dataset of traffic images. It detects vehicles and classifies them into cars, buses, and trucks. A TensorFlow prediction model forecasts traffic congestion trends. The system dynamically adjusts signal timing to optimize flow. 🚀 Future Improvements

🔹 Edge AI Optimization: Improve inference speed using TensorRT. 🔹 Multi-Camera Support: Integrate multiple street cameras for better accuracy. 🔹 Integration with City Infrastructure: Work with smart city APIs for real-world deployment.

👨‍💻 Developer

💡 Omer Kilic – AI & IoT System Development UI/UX, Backend, Data Science

📜 License

This project is licensed under the AGPL-3.0 License.

This README.md gives a comprehensive and structured overview of your project, including setup instructions, technical details, and future improvements. 🚀 Let me know if you'd like any modifications! 🚦