李翠然，谢健骊，高文娟
（兰州交通大学，中国 兰州 730070）
摘要：铁路窄带移动通信系统（GSM-R）正在向铁路宽带移动通信系统（LTE-R）、基于5G的铁路移动通信系统（5G-R）演进。针对未来高速铁路（HSR）通信中的实时视频监控、车-车（T2T）通信、列车多媒体调度等5G-R业务的异构无线网络接入，提出一种基于马尔科夫决策过程（MDP）模型的网络接入算法。根据不同类型业务的服务质量（QoS）属性和无线网络的时变特性构建网络回报函数，并基于模糊层次聚类理论来计算QoS属性的权重值。采用人工智能算法对MDP模型进行求解，使用户以较少的切换次数接入长期期望回报值最大的网络，并仿真分析算法的收敛性和有效性。
关键词：高速铁路；5G-R；异构网络；马尔可夫决策过程；人工智能
Heterogeneous Network Access Technologies Based on 5G-R Services for High-Speed Railway
LI Cuiran, XIE Jianli, GAO Wenjuan
(Lanzhou Jiaotong University, Lanzhou 730070, China)
Abstract: Global system for mobile communications-railway (GSM-R) is evolving to broadband mobile communication systems-R (LTE-R) and 5G for railways (5G-R). A radio access algorithm in wireless heterogeneous environment based on Markov decision process (MDP) decision model is proposed to meet the needs of 5G-R services access to different networks in future high-speed railways (HSRs) communications, such as video surveillance, train-to-train (T2T) direct communication, train multimedia dispatching. According to the quality of service (QoS) attributes of different types of services and time-varying characteristics of wireless networks, the network reward function is constructed, and the QoS attribute weight is determined based on fuzzy clustering theory. The MDP decision model is solved by an artificial intelligence algorithm, which enables users to access the network with the maximum long-term reward with fewer handoffs. In addition, the convergence and effectiveness of the algorithm are analyzed by simulation.
Keywords: high-speed railways; 5G-R; heterogeneous network; Markov decision process; artificial intelligence
.png)
.png)
.png)
