Exploring Traffic Signal Control: A Comprehensive Survey on Reinforcement Learning Techniques

Urban traffic congestion at intersections poses a significant challenge, leading to inefficiencies in traffic flow and increased wait times. Traditional fixed-timer traffic signal systems fail to adapt to fluctuating traffic volumes, exacerbating congestion during peak hours and leading to unnecessary delays during off-peak hours. Reinforcement Learning (RL) offers a promising solution by enabling traffic signal control systems to dynamically adjust signal timings based on real-time data. This paper surveys the current state of traffic signal control systems, highlighting the advantages of RL-based methods and presenting future directions for improvement, including the integration of Internet of Things (IoT) devices, advanced RL algorithms, and enhanced pedestrian detection systems. Through a comparative analysis of existing methods, we propose future enhancements that can further optimize urban traffic management.