AI-Driven Neural Networks for Real-Time Passenger Flow Optimization in High-Speed Rail Networks
Rama Chandra Rao Nampalli, Balaji Adusupalli
Abstract
High-speed rail is now considered the world's most promising mode of travel for medium- and long-distance journeys, and it will become more convenient with the development of technology. The intelligent AI-driven neural networks used in public service facilities, such as railway stations, can be combined with the traffic flow detection system to greatly improve the passenger flow's declaration speed, sensitivity, and accuracy, and to provide more precise and customized passenger service data for passenger flow control and operation management. The study first analyzed the primary factors influencing passenger flow distribution and clearing using the data from Taiyuan South Railway Station and employed a secondary model to simulate and predict the passenger flow distributions in their stations, constructing the neural networks for real-time passenger clearing and then applying it to the station design and management, and summarizing the related recommendations considering passenger flow characteristics and operations control requirements. The proposed study can align the stakeholders' cooperation, promote the balance of the passenger flow between the outstations connecting to each high-speed rail station, and assist the station operator in decision-making that can bring deep benefits.