Traffic signal control for smart cities using reinforcement learning

Abstract

Traffic congestion is increasing globally, and this problem needs to be addressed by the traffic management system. Traffic signal control (TSC) is an effective method among various traffic management systems. In a dynamically changing and interconnected traffic environment, the currently model-based TSCs are not adaptive. In addition, with the rise of smart cities and IoT, there is a need for efficient TSCs that can handle large and complex data. To address this issue, this study proposes a TSC system to maximize the number of vehicles crossing an intersection and balances the signals between roads by using Q-learning (QL). The proposed system has a flexible structure that can be modified to suit the changes in the original structure of the intersection.