基于通信导频数据环境识别研究  

Research on Communication Pilot Data Environment Recognition

作　　者：水宜水 王健 张欣 苏宝睿 易朗宇 朱庆 卢毅 SHUI Yishui;WANG Jian;ZHANG Xin;SU Baorui;YI Langyu;ZHU Qing;LU Yi(The 7th Research Institute of China Electronics Technology Group Corporation,Guangzhou 510310,China)

机构地区：[1]中国电子科技集团公司第七研究所,广东广州510310

出　　处：《移动通信》2024年第9期1-7,共7页Mobile Communications

摘　　要：首先基于真实通信设备,通过对物理层协议进行改造获取用于信道均衡的导频数据,通过对导频数据的数学处理获取信道的时延功率谱。然后对城区建筑物密集的办公区环境、临街道路环境和建筑物较为稀疏的郊区环境进行了实地测量,基于测量得到的时延功率谱数据,通过数据处理得到时延功率谱图片数据,用于对三种环境识别的学习和验证。最后使用深度学习的方式构建了Res Net50深度网络,对数据进行了训练和验证,仿真结果表明,对三种环境的识别准确度可达到90.31%。基于导频数据进行环境识别的研究适用于真实的通信设备进行通感一体化升级,通信设备在无需进行通信协议重构的前提下通过移植训练模型即可实现对环境的识别。Based on the real communication equipment,the pilot data for channel equalization is obtained by modifying the physical layer protocol,and the delay power spectrum of the channel is obtained through the mathematical processing of the pilot data.Then,this paper conducts field measurements of the office environment with dense buildings,the environment of frontage roads and the suburban environment with sparse buildings,and based on the measured delay power spectrum data,the delay power spectrum image data is obtained through data processing for the learning and verification of the three environmental identifications.Finally,the Res Net50 deep network is constructed by deep learning,and the data are trained and verified,and the simulation results show that the recognition accuracy of the three environments can reach 90.31%.In this paper,the research on environment recognition based on pilot data is applicable to the integrated upgrade of real communication equipment,which can realize the recognition of the environment by transplanting the training model without the need for communication protocol reconstruction.

关 键 词：深度学习 导频 时延功率谱 环境识别 

分 类 号：TN929.5[电子电信—通信与信息系统]