A Survey on AI-Based Adaptive Traffic Signal Control Systems for Smart Cities

Traffic congestion has become a serious problem in urban areas due to the rapid increase in the 
number of vehicles and the limitations of traditional fixed-time traffic signal systems. These systems 
are not able to adjust according to real-time traffic conditions, which leads to delays, higher fuel 
consumption, and increased environmental pollution [20]. To overcome these issues, recent research 
focuses on developing intelligent and adaptive traffic signal control systems using technologies such 
as Artificial Intelligence, Machine Learning, Deep Learning, and Reinforcement Learning [9]. Computer 
vision techniques like YOLO are commonly used for real-time vehicle detection and traffic density 
estimation [4], while IoT and edge computing help in efficient data collection and faster processing [1], 
[16]. 
This survey paper reviews recent work (2025–2026) in the field of smart traffic management. It also 
looks at systems designed for emergency vehicle prioritization and traffic forecasting using predictive 
analytics [3], [11]. Different approaches are compared based on their efficiency, scalability, and real
time performance. From the study, it can be observed that AI-based adaptive systems improve traffic flow and reduce congestion effectively. However, their practical implementation on a large scale is still 
challenging due to high cost and system complexity [5], [14]. Future research can focus on making 
these systems more affordable, scalable, and suitable for real-world smart city applications.