Skip to main content
ESTIN

Ramp Metering Control

Optimized highway traffic flow with a Q-learning algorithm, managing traffic lights, lane changes, and car following for improved efficiency.
Project image

For this class project, I designed and implemented a Q-learning algorithm to optimize highway traffic management, addressing challenges like congestion and inefficiency in traffic flow.

  1. Problem Analysis: Studied real-world traffic patterns and identified key factors such as ramp metering, traffic light control, and lane changes that influence traffic efficiency.

  2. Reinforcement Learning Model:

    • Developed a Q-learning model to dynamically manage ramp meters and traffic signals.

    • Incorporated factors like car-following behavior and lane-changing dynamics into the reward function to encourage efficient traffic flow.

  3. Simulation: Conducted simulations using Python to evaluate the model’s performance in reducing congestion and improving throughput.

  4. Results: The model demonstrated significant improvements in traffic flow efficiency, showcasing the effectiveness of RL in solving transportation challenges.

#