基于GAN的铁路5G-R车地无线通信测试数据生成方法
Test Data Generation Method of Railway 5G-R Train-trackside Wireless Communication Based on GAN
魏斌 1邹劲柏 1刘立海 2王耀国 2陈砚明3
作者信息
- 1. 上海应用技术大学轨道交通学院,上海 201418
- 2. 中铁第四勘察设计院集团有限公司,湖北武汉 430063
- 3. 包神铁路集团有限公司,内蒙古包头 014010
- 折叠
摘要
鉴于当前铁路5G-R应用研究中通信测试数据少的问题,结合公网5G基站的有关标准和铁路无线通信业务应用特点,提出面向铁路通信场景的基于GAN的数据生成方法.通过介绍GAN算法原理,分析5G-R通信数据特征,采用GAN模型进行5G-R业务数据生成仿真,并设计LSTM预测模型验证生成样本真实度与3GPP标准及相关规范的符合性.实验结果表明:5G-R车地无线通信测试数据GAN模型训练结果可靠,可用于5G-R业务负荷相关应用研究.
Abstract
Targeting at the insufficiency of communication test data in the current railway 5G-R application research,this paper proposed a GAN-based data generation method for railway communication scenarios,taking account of the relevant standards of public network 5G base stations and the characteristics of railway wireless communication service applications.By introducing the principle of GAN algorithm,this paper analyzed the characteristics of 5G-R communication data,used the GAN model to simulate the generation of 5G-R service data,and designed an LSTM prediction model to verify the compliance of the authenticity of generated samples with 3GPP standards and relevant specifications.The experimental results showed that the GAN model training result of 5G-R train-trackside wireless communication test data is reliable and can be used for application research related to 5G-R service load.
关键词
5G-R/GAN/数据生成/LSTM/车地无线通信Key words
5G-R/GAN/data generation/LSTM/train-trackside wireless communication引用本文复制引用
基金项目
中国国家铁路集团有限公司科技研究开发计划项目(N2022G048)
出版年
2024
NETL
NSTL
万方数据