大规模MIMO系统基于多分辨率深度学习网络的CSI反馈研究  被引量:2

Research on CSI feedback of massive MIMO system based on multi-resolution deep learning network

在线阅读下载全文

作  者:李中捷[1] 熊吉源 高伟 金闪 LI Zhongjie;XIONG Jiyuan;GAO Wei;JIN Shan(Hubei Key Laboratory of Intelligent Wireless Communications & College of Electronic andInformation Engineering,South-Central University for Nationalities,Wuhan 430074,China)

机构地区:[1]中南民族大学电子信息工程学院&智能无线通信湖北重点实验室,武汉430074

出  处:《中南民族大学学报(自然科学版)》2021年第1期50-56,共7页Journal of South-Central University for Nationalities:Natural Science Edition

基  金:国家自然科学基金资助项目(61379028,61671483);湖北省自然科学基金资助项目(2016CFA089);中央高校基本科研业务费专项资金资助项目(CZY19003)。

摘  要:快变信道环境下,采用频分双工模式下的大规模MIMO系统,用户通过反馈链路将信道状态信息(Channel State Information,CSI)发送给基站,为适应信道快速变化保证系统性能,要求降低反馈时延及减少反馈开销.提出一种基于深度学习的多分辨率信道状态信息网络(Multi-resolution Channel State Information Network,MCSINet),对反馈的信道状态信息进行压缩及预测,能够显著减少信道状态信息捕获与反馈开销,及降低时延.MCSINet模拟信道状态信息编解码系统,采用残差网络从信道样本中学习并完成信道预测,并通过多分辨率的卷积操作以及针对不同压缩率改变网络结构,从而更好预测信道状态.实验结果表明:与LASSO,TVAL3,CSINet等方法相比,MCSINet可以显著提高恢复信道状态信息,并且具有更低的误码率,复杂度和时延.In a fast time-varying channel environment and a massive MIMO system using frequency division duplex mode,users send channel state information(Channel State Information,CSI)to the base station through the feedback link,to adapt to the rapid channel changes to ensure system performance,reduce requirements feedback delay and reduce feedback overhead.A deep learning-based multi-resolution channel state information network(Multi-resolution Channel State Information Network,MCSINet)is proposed to compress and predict the feedback channel state information,to significantly reduce the cost of channel state information capture and feedback,and reduce the purpose of the delay.MCSINet simulates the channel state information coding and decoding system,using the residual network to learn from channel samples and complete channel prediction,through multi-resolution convolution operation and changing the network structure for different compression rates,to predict the channel state better.Compared with LASSO,TVAL3,CSINet,and other methods,MCSINet can significantly improve the recovery of channel state information and has a lower bit error rate,complexity,and delay.

关 键 词:大规模MIMO 信道状态信息反馈 深度学习 

分 类 号:TN929.5[电子电信—通信与信息系统]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象