| 基于BP和RBF神经网络的C-V2X无线资源管理 |
| |
| 引用本文： | 冯毅,葛宁,张陶冶.基于BP和RBF神经网络的C-V2X无线资源管理[J].电讯技术,2023,63(11):1651-1660. |
| |
| 作者姓名： | 冯毅 葛宁 张陶冶 |
| |
| 作者单位： | 1.清华大学 电子工程系，北京 100091；2.中国联合网络通信有限公司智网创新中心，北京 100037 |
| |
| 基金项目： | 清华大学-中国移动通信集团有限公司联合研究院项目（20212930006）；南通大学智能信息技术研究中心开放课题项目（20192001872） |
| |
| 摘 要： | 为了提升蜂窝车联网（Cellular Vehicle-to-Everything,C-V2X）资源复用的有效性和降低终端间的干扰，提出通过神经网络对未来时刻车流量的预测辅助无线资源管理方案。依据车载单元（On Board Unit,OBU）与路侧单元（Road Side Unit,RSU）间的车联网消息，获取RSU覆盖区域内各时刻的车流情况，分别采用BP（Back Propagation）神经网络和RBF（Radial Basis Function）神经网络进行短时交通流预测。RSU根据预测结果进行自适应分簇，簇间复用相同资源，簇内进行资源池的划分，RSU覆盖内的OBU在划分的资源池中选择发送资源，从而减少终端间的干扰，并保证热点区域车辆拥有更多的资源。仿真结果表明，在道路交通拥塞的场景下，所提方案的数据包接收率较标准中的方案提升14%，较典型文献方案提升10%，保证了通信的可靠性。
|
| 关 键 词： | 蜂窝车联网（C-V2X） 无线资源管理 神经网络 资源池 |
BP and RBF Neural Network Based Radio Resource Management in C-V2X Networks |
| |
| FENG Yi,GE Ning,ZHANG Taoye.BP and RBF Neural Network Based Radio Resource Management in C-V2X Networks[J].Telecommunication Engineering,2023,63(11):1651-1660. |
| |
| Authors: | FENG Yi GE Ning ZHANG Taoye |
| |
| Affiliation: | 1.Department of Electronic Engineering,Tsinghua University,Beijing 100091,China;2.China United Network Communication Co.,Ltd.Intelligent Network Innovation Center,Beijing 100037,China |
| |
| Abstract: | In order to improve the effectiveness of Cellular Vehicle-to-Everything(C-V2X) resource reuse and reduce the interference between terminals,the authors propose a scheme to improve radio resource management by predicting the traffic flow through neural networks.The traffic flow in the road side unit(RSU) coverage area is obtained through V2X messages between on board unit(OBU) and RSU.Back Propagation(BP) neural network and Radial Basis Function(RBF) neural network are used for short-time traffic flow prediction respectively.RSUs perform adaptive clustering according to the prediction results.The resource pools are multiplexed between clusters and divided within the clusters.OBUs within RSU coverage select transmitting resources in the divided resource pool,thus reducing interference between terminals and ensuring more resources for vehicles in hotspot areas.Simulation results show that the packet reception ratio of the proposed scheme improves by 14% compared with the scheme in the specification and 10% compared with the typical literature scheme in a traffic congestion scenario,ensuring the reliability of communication. |
| |
| Keywords: | cellular Vehicle-to-Everything(C-V2X) radio resource management neural network resource pool |
|
| 点击此处可从《电讯技术》浏览原始摘要信息 |
|
点击此处可从《电讯技术》下载全文 |
