Trafficient

An intelligent traffic control system using Deep Q-Learning to optimize traffic signal timing. This project achieved a 50% reduction in average wait times and a 34.9% increase in traffic flow speed.

Overview

Trafficient uses reinforcement learning to dynamically control traffic signals, adapting to real-time traffic conditions. The system is built on SUMO (Simulation of Urban MObility) and PyTorch, implementing a Deep Q-Network (DQN) for intelligent decision-making.

Key Features

• Real-time traffic signal optimization
• Predictive queue management
• Dynamic phase timing
• Multi-intersection support
• Performance metrics tracking

Requirements

• Python 3.8+
• SUMO (Simulation of Urban MObility)
• PyTorch
• NumPy
• Matplotlib

Installation

1. Install SUMO:

# Set SUMO_HOME environment variable
export SUMO_HOME="path/to/sumo"

2. Install Python dependencies:

pip install torch numpy matplotlib

3. Clone the repository:

git clone [repository-url]
cd trafficient

Usage

1. Run the reinforcement learning model:

python rl_traffic_control.py

2. Compare with traditional methods:

python compare_methods.py

3. Generate random traffic patterns:

python random_trips.py

Project Structure

trafficient/
├── rl_traffic_control.py     # Main RL implementation
├── traditional_traffic.net.xml    # Network configuration
├── traditional_traffic.sumo.cfg   # SUMO configuration
├── random_trips.py           # Traffic generation
├── traditional_traffic_control.py # Traditional Traffic control
├── random_traffic.rou.xml # Traditional Traffic control
└── compare_methods.py        # Performance comparison

Results

• Average wait time reduction: 50%
• Traffic flow speed increase: 34.9%
• Improved throughput across 8+ traffic approaches

Acknowledgments

• SUMO Traffic Simulation
• PyTorch Team
• Traffic Control Research Community