摘要
针对5G网络规划与优化存在的问题,提出一种融合系统仿真和深度神经网络模型的网络时延预测方法.基于射线追踪模型、高清地图、工程参数等构建时延仿真模型,利用时延仿真模型获取大量时延数据.基于无线通信理论提出三视图特征模型,此模型用于输入特征提取.通过深度神经网络学习时延数据特征,训练神经网络模型,利用神经网络模型预测网络时延.实验结果表明该方法具有可行性和有效性.
Aiming at the problems of 5G network planning and optimization,a delay prediction method that combined system simulation and deep neural network models was proposed.Based on ray tracing models,high-definition maps,engineering parameters,etc.,a delay simulation model was built to obtain a large amount of delay data.Based on wireless communication theory,a three-view feature model was proposed.This model was used to input feature extraction.The characteristics of delay data were learnt through deep neural networks.The neural network models were trained.The network delays were predicted by neural network models.Experiment results showed that the method was feasible and effective.
作者
朱军
臧守涛
李剑
李汐
叶国骏
ZHU Jun;ZANG Shoutao;LI Jian;LI Xi;YE Guojun(School of Electronics and Information Engineering,Anhui University,Hefei 230601,China;Shanghai Institute of Microsystem and Information Technology,Chinese Academy of Sciences,Shanghai 200050,China;Huawei Technologies Co.,Ltd.,Shanghai 201206,China)
出处
《安徽大学学报（自然科学版）》
CAS
北大核心
2020年第2期35-42,共8页
Journal of Anhui University(Natural Science Edition)
基金
安徽省科技重大专项(18030901010)
教育部产学合作协同育人项目(201801129061)。
关键词
5G
网络规划与优化
射线追踪模型
深度学习
时延预测
5G
network planning and optimization
ray tracing model
deep learning
delay prediction
