Research On Network Slicing Resource Management In 5G-based Urban Railway Transit Train-ground Communication System

At present,urban rail transit is developing in the direction of automation and intelligence.Vehicle-ground communication services are more diverse,and higher requirements are placed on wireless communication systems.For example,train video surveillance requires higher bandwidth,and the next-generation train control system requires much lower latency.Currently,the long term evolution for metro(LTE-M),as the mainstream wireless system for vehicle-ground communication,has limitations in bandwidth and latency,and cannot fully meet the needs of future vehicle-ground communication systems.Compared with LTE-M,the performance of the fifth-generation mobile communication technology has been significantly improved in all aspects,and 5G network slicing technology can divide different logically isolated networks to meet the needs of various vehicle-to-ground wireless services with different requirements,which is more secure and suitable for management.Based on the development trend of rail transit vehicle-ground communication system,this thesis analyzes the future vehicle-ground wireless services’ requirements,researches and designs how to build a vehicle-ground communication system based on5 G network slicing,and according to the characteristics of rail transit and end-to-end network slicing,improves the performance of the vehicle-ground communication system in radio access network and core network with virtual computing resources,and a semi-physical slicing platform is build.The main work is summarized as follows:(1)The development of urban rail transit vehicle-ground communication system and the demand for 5G technology are summarized,and the related technologies of 5G network slicing,the research status of 5G-based rail transit communication system and the research status of network slicing resource management are introduced.In view of the needs and development trends of vehicle-ground communication services,various plans for rail transit data communication system based on public and private 5G networks are provided.It is made clear that network slicing is the key technology for the future vehicle-ground communication system to meet various types of service bearers,and it is necessary to study the management of sliced radio resources and virtualized resources to improve the performance of the communication system.(2)Combined with the characteristics of the vehicle-ground communication system environment and the different Qo S requirements of services,a network slicing radio resource allocation scheme based on service priority and a dynamic slicing radio resource allocation algorithm based on long short-term memory network and deep reinforcement learning are proposed.The simulation verifies that the designed algorithm can effectively improve the service quality of each network slice of vehicle-ground communication system.(3)Aiming at the requirement for prioritized guarantee of security services and high computing resources of services such as video surveillance and passenger communication services in vehicle-ground communication system,the computing resource allocation problem of vehicle-ground communication based on network slicing and edge computing is modeled,and a priority-based heuristic slicing computing resource allocation algorithm(PHSA)is designed.Through the simulation verification,the effectiveness of the algorithm is illustrated from four aspects such as computing resource utilization.(4)Based on the open source 5G software radio Open Air Interface(OAI),the half-physical slicing simulation platform is built with the customized vehicle-ground communication network slicing architecture,and three kinds of slicing networks are established and service bearer tests are carried out,which verifies the feasibility of 5G network slicing in bearing vehicle-ground wireless services and the isolation between network slices.There are 48 figures,12 tables and 85 references.