AI-Traffic-Dhaka

AI-Powered Dynamic Traffic Management System for Dhaka City 🇧🇩

This project presents an AI-based traffic control system using Reinforcement Learning to optimize vehicle flow and reduce congestion in Dhaka, Bangladesh. The system uses the SUMO simulator and a Deep Q-Network (DQN) agent to dynamically adjust traffic signals across multiple intersections.

🚦 Key Features

• Multi-intersection traffic control
• SUMO simulation of Dhaka traffic
• Deep Q-Network (DQN) implementation
• Up to 35% reduction in wait time vs static signals

🧠 Technologies Used

• Python
• SUMO (Simulation of Urban MObility)
• TraCI
• TensorFlow / PyTorch
• Git & Trello (Project Management)

📂 Project Structure

• /code: DQN agent, training script, environment wrappers
• /simulation: SUMO config files, network map, vehicle types
• /results: Output logs and performance metrics
• /docs: Final report, diagrams
• video_demo.mp4: Walkthrough and results

🎓 Academic Context

This was my final year project at Ulster University for the COM668 Computing Project module.
Mentor: Baba Shaheer | Student ID: B00916950

📽️ Video Demo

Watch a short walkthrough of the simulation and results:
[Upload video_demo.mp4 here or host on YouTube and link it]

📈 Results

• Up to 35% reduction in average wait time.
• Improved throughput and reduced queue lengths at critical intersections.

📦 Setup Instructions

```bash git clone https://github.com/your-username/AI-Traffic-Dhaka cd AI-Traffic-Dhaka pip install -r requirements.txt

This site is open source. Improve this page.