Performance Evaluation of a 5G-Powered Vehicle-to-Infrastructure (V2I) Framework for Intelligent Traffic ControlOpen Access

Abstract

Urban mobility has become a major concern due to urban sprawl, increased traffic and difficulties in improving infrastructure. Congestion, increased carbon emissions and fuel waste result from the incapacity of current traffic management systems to react swiftly and effectively. To solve this issue, we propose creating an Intelligent Urban Traffic Management System (IUTMS) that connects vehicles and infrastructure to improve traffic flow. This system would combine 5G technology, Internet of Things (IoT) sensors, and cloud-based artificial intelligence (AI) analytics. The system incorporates a variety of Internet of Things devices, including computer vision cameras, LiDAR modules, and infrared sensors. These devices monitor vehicles and their surroundings. The devices transmit this data to a cloud-based service for processing and decision-making via low-latency 5G wireless communication. Our IUTMS system was tested in an 8x8-kilometre urban study area with 36 junctions, powered by Apache Kafka, ns-3 (5G-LENA) and SUMO. The proposed framework has an end-to-end latency of between 6.87 and 15.94 milliseconds. Even under truck-density stress, the service remains reliable, achieving a Packet Delivery Ratio (PDR) of 85.46% to 95.86%, a Network Throughput of 78.12% to 93.84% and latency of less than 10 ms. This work improves and validates the prediction and input aspects of the suggested framework, leading to a low-latency, high-reliability and efficient real-world system.