Synchronization of traffic lights in smart cities

Abstract

Nowadays, STLSs (Smart Traffic Light Systems) are widely adopted by smart cities to control traffic lights. They often rely on dedicated equipment like camera and sensors to collect traffic data. However, STLSs in urban areas are not flexible enough to efficiently manage traffic tricky problems such , emergencies, etc. Therefore, the increasing need of automatic synchronized STLS system of various traffic controllers at different intersections which performed with reduced occupancy and waiting time benefits becomes a major concern. This is crucial for meeting the optimized traffic parameters of multiple intersections and driver's needs. Our objective is to develop a new system, called ADSTLS (Adaptive and Dynamic Smart Traffic Light System) to address traffic management at an intersection, solving the challenging problem of traffic congestion while prioritizing emergency vehicles. Therefore, we have proposed a system with new hybrid traffic flow model that combines a cycle model and a phase model for optimizing and efficiently managing traffic light planning, along with a decision-making approach focused on reducing congestion and average vehicle waiting time. By collecting traffic data, the system automatically extracts useful traffic information using computer vision and computing standard traffic metrics. Moreover, we proposed two-traffic modes for regular and emergency vehicles to achieve an optimal decision-making process. The first mode, the dynamic mode, select the best phase using the Weight Chicken Swarm Optimization (WCSO) algorithm to ensure an optimal vehicle waiting time and queue occupancy at the city’s intersection. The second mode, the adaptive mode, determines the priority direction based on the distance of emergency vehicles and their priority levels. We have also proposed extending ADSTLS system using the multi-agent paradigm to improve the system’s performance in terms of execution time. To demonstrate our approach, we have presented a simulator applied to a real case study of EL‐Hidhab Setif city intersection. The experimental results showed a decrease in the average vehicle waiting time (31 s) and queue occupation rate (33.82%) across all simulated traffic scenarios. Furthermore, compared to other car types, emergency vehicles usually had much shorter wait times.