Artificial Intelligence-Based Real-Time Traffic Signal Optimization for Smart City Infrastructure Integration

As with any other major city, Fresno faces significant challenges with traffic congestion during rush hours, causing difficulties for commuters. This also causes substantial emission problems. This proof-of-concept (POC) research project aims to upgrade the street traffic signal by deploying an artificial intelligence algorithm called deep reinforcement learning (DRL). The deep reinforcement learning algorithm provides the core software module that will be hosted in the adaptive traffic management system, where the system can dynamically adjust to changing traffic patterns, reduce wait times, and minimize environmental impact. To develop the algorithm, much high-level coding work is expected and a computer virtual simulation environment called Simulation of Urban MObility (SUMO). Referencing earlier work by other researchers, this project aims to deliver a Fresno-tailored traffic management system. The proposed solution will not only address Fresno's specific urban transportation challenges but also provide a scalable framework that other midsized cities can adopt.