当前位置>主页 > 期刊在线 > 通信工程 >

通信工程21年6期

基于卷积神经网络的短波信道质量分类
孙汝杰
(江苏省交通技师学院,江苏 镇江 212028)

摘  要:自适应短波通信系统可以解决短波信道质量差、频率资源短缺等问题,而信道质量估计是其中的重要环节。为了避免基于深度学习的传统方法中基带信号过大而无法提取的问题,该文将基带信号转换成星座轨迹图,再分别采用 AlexNet,ResNet 和 DenseNet 三种卷积神经网络对其进行训练。实验结果验证了该文提出方法的可行性,且随着网络的加深,准确度也将提升。


关键词:自适应短波通信;卷积神经网络;星座轨迹图;深度学习;信道质量分类



DOI:10.19850/j.cnki.2096-4706.2021.06.018


中图分类号:TP183;TN925                          文献标识码:A                                    文章编号:2096-4706(2021)06-0070-04


HF Channel Quality Classification Based on Convolutional Neural Network

SUN Rujie

(Jiangsu Jiaotong College,Zhenjiang 212028,China)

Abstract:Adaptive short-wave communication system can solve the problems of poor quality of short-wave channel and frequency resource shortage,while channel quality estimation is an important they are combined to simulate communication simulation. among them. In order to avoid the problem that the baseband signal is too large to be extracted in the traditional method based on deep learning, this paper transforms the baseband signal into constellation trajectories diagram,and then uses three convolutional neural networks of AlexNet,ResNet and DenseNet for training it. The experimental results verify the feasibility of the proposed method in this paper. With the deepening of the network,the accuracy will also be improved.

Keywords:adaptive short-wave communication;convolutional neural network;constellation trajectories diagram;deep learning;channel quality classification


参考文献:

[1] 左卫 . 短波通信系统发展及关键技术综述 [J]. 通信技术, 2014,47(8):847-853.

[2] 郭英波 . 短波通信中的信道估计与同步技术研究 [D]. 北 京:北京邮电大学,2015.

[3] 赵兰浩.短波通信技术发展综述 [J].信息通信,2017(11): 257-258.

[4] 王高峰 . 短波通信系统自适应信道均衡技术应用研究 [J]. 中国新通信,2018,20(5):98.

[5] 廖勇,花远肖,姚海梅 . 基于深度学习的 OFDM 信道估计 [J]. 重庆邮电大学学报(自然科学版),2019,31(3):348-353.

[6] 李坤,张静,李潇,等 . 人工智能辅助的信道估计最新研 究进展 [J]. 电信科学,2020,36(10):46-55.


作者简介:孙汝杰(1984.04—),男,汉族,江苏如皋人, 工程师,信息管理系实训中心主任,硕士研究生,研究方向:信息 工程、人工智能与通信技术等。