the project aims to use reinforcement learning (Deep Q-learning) to optimize traffic signal timings at a specific intersection in Morocco. We used SUMO (Simulation of Urban MObility) to simulate the real intersection and gather virtual data on traffic flow. Then We used this data to train the reinforcement learning algorithm to optimize traffic signal timings in real-time, with the goal of minimizing congestion and improving traffic flow. The goal is to improve the traffic flow using the simulation environment generated by SUMO and RL techniques.
this project is based on the method presented in the article "A Deep Reinforcement Learning Approach to Adaptive Traffic Lights Management" which is written by Andrea Vidali, Luca Crociani, Giuseppe Vizzari and Stefania Bandini.