基于深度学习的大规模MIMO信道状态信息反馈

doi:10.11959/j.issn.2096-3750.2020.00157

物联网学报 ›› 2020, Vol. 4 ›› Issue (1): 33-44.doi: 10.11959/j.issn.2096-3750.2020.00157

基于深度学习的大规模MIMO信道状态信息反馈

陈慕涵,郭佳佳,李潇,金石()

东南大学移动通信国家重点实验室，江苏南京 210096

修回日期:2020-03-05 出版日期:2020-03-30 发布日期:2020-03-28
作者简介:陈慕涵（1997- ），女，江苏南通人，东南大学移动通信国家重点实验室硕士生，主要研究方向为基于深度学习的信道状态信息反馈、机器学习等|郭佳佳（1993- ），男，江苏泰兴人，东南大学移动通信国家重点实验室博士生，主要研究方向为基于深度学习的信道状态信息反馈等|李潇（1982- ），女，安徽蚌埠人，博士，东南大学移动通信国家重点实验室副教授、硕士生导师，主要研究方向为移动通信理论与关键技术、智能通信以及智能反射表面在无线通信中的应用等|金石（1974- ），男，安徽黄山人，博士，东南大学移动通信国家重点实验室教授、博士生导师，主要研究方向为移动通信理论与关键技术、物联网理论与关键技术以及人工智能在无线通信中的应用等
基金资助:
国家自然科学基金资助项目(61941104)

An overview of the CSI feedback based on deep learning for massive MIMO systems

Muhan CHEN,Jiajia GUO,Xiao LI,Shi JIN()

National Mobile Communications Research Laboratory,Southeast University,Nanjing 210096,China

Revised:2020-03-05 Online:2020-03-30 Published:2020-03-28
Supported by:
The National Science Foundation of China(61941104)

摘要/Abstract

摘要：

大规模多输入多输出（MIMO,multiple-input multiple output）技术被认为是下一代移动通信的核心技术之一，其系统增益建立在基站能够精确获知信道状态信息（CSI,channel state information）的基础上。由于天线数量显著增长，传统基于码本或矢量量化的反馈方案面临较大的技术挑战，而深度学习（DL,deep learning）为解决大规模MIMO系统的CSI反馈问题提供了新思路。围绕大规模MIMO系统CSI反馈关键技术展开调研，首先阐述了CSI反馈的研究背景和意义，接着构建大规模MIMO系统模型并分析CSI的稀疏特性，然后详细介绍和比较了国内外将DL技术引入CSI反馈机制中的方案，最后对基于DL的CSI反馈的未来发展趋势做了进一步展望。

关键词: 大规模MIMO, 深度学习, CSI反馈

Abstract:

The massive multiple-input multiple-output (MIMO) technology is considered to be one of the core technologies of the next generation communication system.To fully utilize the potential gains of MIMO systems,the base station should accurately acquire the channel state information (CSI).Due to the significant increase in the number of antennas,the traditional feedback schemes based on the codebook or vector quantization are faced with great technical challenges.Recently,deep learning (DL) has provided a new idea for solving CSI feedback problems in massive MIMO systems.It was focused on the key technologies of the CSI feedback for massive MIMO systems.Firstly,the background and significance of the CSI feedback were expounded.Then,a model for the massive MIMO system was established and the sparse nature of CSI was analyzed.Several schemes of introducing DL into the CSI feedback mechanism were introduced and compared in detail.Finally,a further prospect on the development trend of the CSI feedback based on DL was made.

Key words: massive MIMO, deep learning, CSI feedback

中图分类号:

TN929.5

陈慕涵,郭佳佳,李潇,金石. 基于深度学习的大规模MIMO信道状态信息反馈[J]. 物联网学报, 2020, 4(1): 33-44.

Muhan CHEN,Jiajia GUO,Xiao LI,Shi JIN. An overview of the CSI feedback based on deep learning for massive MIMO systems[J]. Chinese Journal on Internet of Things, 2020, 4(1): 33-44.

图/表 12

图1

表1

图2

表2

图3

表3

图4

图5

图6

图7

表4

图8

参考文献 29

[1]	WONG V S , SCHOBER R , NG D W K ,et al. Key technologies for 5G wireless systems[M]. Cambridge: Cambridge University PressPress, 2017.
[2]	YANG P , XIAO Y , XIAO M ,et al. 6G wireless communications:vision and potential techniques[J]. IEEE Network, 2019,33(4): 70-75.
[3]	LOVE D J , HEATH R W , LAU V K N ,et al. An overview of limited feedback in wireless communication systems[J]. IEEE Journal on Selected Areas in Communications, 2008,26(8): 1341-1365.
[4]	ZHOU Y , HERDIN M , SAYEED A M ,et al. Experimental study of MIMO channel statistics and capacity via the virtual channel representation[D]. Madison:University of Wisconsin-Madison, 2007.
[5]	KYRITSI P , COX D C , VALENZUELA R A ,et al. Correlation analysis based on MIMO channel measurements in an indoor environment[J]. IEEE Journal on Selected Areas in Communications, 2003,21(5): 713-720.
[6]	KUO P H , KUNG H T , TING P A . Compressive sensing based channel feedback protocols for spatially-correlated massive antenna arrays[C]// 2012 IEEE Wireless Communications and Networking Conference (WCNC). IEEE, 2012: 492-497.
[7]	RAO X B , LAU V K . Distributed compressive CSIT estimation and feedback for FDD multi-user massive MIMO systems[J]. IEEE Transactions on Signal Processing, 2014,62(12): 3261-3271.
[8]	SHEA T O , HOYDIS J . An introduction to deep learning for the physical layer[J]. IEEE Transactions on Cognitive Communications and Networking, 2017,3(4): 563-575.
[9]	QIN Z J , YE H , LI G Y ,et al. Deep learning in physical layer communications[J]. IEEE Wireless Communications, 2019,26(2): 93-99.
[10]	张静, 金石, 温朝凯 ,等. 基于人工智能的无线传输技术最新研究进展[J]. 电信科学, 2018,34(8): 46-55.
	ZHANG J , JIN S , WEN C K ,et al. An overview of wireless transmission technology utilizing artificial intelligence[J]. Telecommunications Science, 2018,34(8): 46-55.
[11]	WANG T Q , WEN C K , WANG H Q ,et al. Deep learning for wireless physical layer:opportunities and challenges[J]. China Communications, 2017,14(11): 92-111.
[12]	WEN C K , JIN S , WONG K K ,et al. Channel estimation for massive MIMO using Gaussian-mixture Bayesian learning[J]. IEEE Transactions on Wireless Communications, 2015,14(3): 1356-1368.
[13]	DAUBECHIES I , DEFRISE M , MOL C D . An iterative thresholding algorithm for linear inverse problems with a sparsity constraint[J]. Communications on Pure and Applied Mathematics, 2003,57(11): 1413-1457.
[14]	DONOHO D L , MALEKI A , MONTANARI A . Message passing algorithms for compressed sensing[J]. Proceedings of the National Academy of Sciences of the United States of America, 2009,106(45): 18914-18919.
[15]	LI C , YIN W , ZHANG Y ,et al. TVAL3:TV minimization by augmented Lagrangian and alternating direction algorithms[E/OL]. .
[16]	METZLER C A , MALEKI A , BARANIUK R G . From denoising to compressed sensing[J]. IEEE Transactions on Information Theory, 2016,62(9): 5117-5144.
[17]	LOHIT S , KULKARNI K , KERVICHE R ,et al. Convolutional neural networks for non-iterative reconstruction of compressively sensed images[J]. IEEE Transactions on Computational Imaging, 2018,4(3): 326-340.
[18]	MOUSAVI A , DASARATHY G , BARANIUK R G . DeepCodec:adaptive sensing and recovery via deep convolutional neural networks[J]. arXiv:1707.03386, 2017
[19]	WEN C K , SHIH W T , JIN S . Deep learning for massive MIMO CSI feedback[J]. IEEE Wireless Communications Letters, 2018,7(5): 748-751.
[20]	VINCENT P , LAROCHELLE H , BENGIO Y ,et al. Extracting and composing robust features with denoising autoencoders[C]// Proceedings of the 25th International Conference on Machine Learning (ICML). ACM, 2008: 1096-1103.
[21]	LIU L F , OESTGES C , POUTANEN J ,et al. The COST 2100 MIMO channel model[J]. IEEE Wireless Communications, 2012,19(6): 92-99.
[22]	XU K , REN F B . CSVideoNet:a real-time end-to-end learning framework for high-frame-rate video compressive sensing[C]// 2018 IEEE Winter Conference on Applications of Computer Vision (WACV). IEEE, 2018: 1680-1688.
[23]	WANG T Q , WEN C K , JIN S ,et al. Deep learning-based CSI feedback approach for time-varying massive MIMO channels[J]. IEEE Wireless Communications Letters, 2019,8(2): 416-419.
[24]	LU C , XU W , SHEN H ,et al. MIMO channel information feedback using deep recurrent network[J]. IEEE Communications Letters, 2019,23(1): 188-191.
[25]	LIU Z Y , ZHANG L , DING Z . Exploiting bi-directional channel reciprocity in deep learning for low rate massive MIMO CSI feedback[J]. IEEE Wireless Communications Letters, 2019,8(3): 889-892.
[26]	WU P X , LIU Z C , CHENG J L . Compressed CSI feedback with learned measurement matrix for mmWave massive MIMO[J]. arXiv:1903.02127, 2019
[27]	LI X Y , WU H M . Spatio-temporal representation with deep neural recurrent network in MIMO CSI feedback[J]. IEEE Wireless Communications Letters, 2020
[28]	CAI Q Y , DONG C , NIU K . Attention model for massive MIMO CSI compression feedback and recovery[C]// 2019 IEEE Wireless Communications and Networking Conference (WCNC). IEEE, 2019.
[29]	GUO J J , WANG J H , WEN C K ,et al. Compression and acceleration of neural networks for communications[J]. arXiv:1907.13269, 2019

场景 γ	LASSO	TVAL3	BM3D-AMP	CsiNet
1/4	-7.59	-14.87	-4.33	-17.36
室内1/16	-2.72	-2.61	0.26	-8.65
1/32	-1.03	-0.27	24.72	-6.24
1/64	-0.14	0.63	26.22	-5.84
1/4	-5.08	-6.90	-1.33	-8.75
室外1/16	-1.01	-0.43	0.55	-4.51
1/32	-0.24	0.46	22.66	-2.81
1/64	-0.06	0.76	25.45	-1.93

场景		γ	LASSO	TVAL3	BM3D-AMP	CsiNet	CsiNet-LSTM
		1/16	-2.96	-3.20	0.25	-10.59	-23.06
室内	NMSE/dB	1/32	-1.18	-0.46	20.85	-7.35	-22.33
		1/64	-0.18	0.60	26.66	-6.09	-21.24
		1/16	0.247 1	0.314 8	0.345 4	0.000 1	0.000 3
	运行时间/s	1/32	0.213 7	0.314 8	0.555 6	0.000 1	0.000 3
		1/64	0.247 9	0.286 0	0.604 7	0.000 1	0.000 3
		1/16	-1.09	-0.53	0.40	-3.60	-9.86
室外	NMSE/dB	1/32	-0.27	0.42	18.99	-2.14	-9.18
		1/64	-0.06	0.74	24.42	-1.65	-8.83
		1/16	0.212 2	0.314 5	0.421 0	0.000 1	0.000 3
	运行时间/s	1/32	0.240 9	0.298 5	0.603 1	0.000 1	0.000 3
		1/64	0.016 6	0.285 0	0.598 0	0.000 1	0.000 3

方法	γ
方法	1/16	1/32	1/64
CsiNet	530 656	268 448	137 344
CsiNet-LSTM	102 009 892	101 354 404	101 026 660
RecCsiNet	19 478 584	18 118 392	17 450 584
PR-RecCsiNet	793 144	333 816	153 304

测量矩阵	m=20	m=25	m=30	m=35	m=40
Φ?	95.90%	98.70%	99.60%	100%	100%
F	0.90%	7.85%	89.20%	99.80%	99.75%
S	5.30%	30.15%	72.70%	90.00%	98.45%
B	5.90%	26.80%	63.10%	87.70%	99.10%
G	2.15%	13.45%	58.50%	84.75%	97.70%
P	0	0	0.45%	1.00%	6.85%

基于深度学习的大规模MIMO信道状态信息反馈

An overview of the CSI feedback based on deep learning for massive MIMO systems

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 29

相关文章 11

Metrics

推荐阅读 0

[1]	张志飞, 刘峰, 葛祎阳, 李烁, 张煜, 熊轲. 一种基于深度可分离卷积和注意力机制的入侵检测方法[J]. 物联网学报, 2023, 7(1): 49-59.
[2]	蒋锐, 孙刘婷, 王小明, 李大鹏, 徐友云. 基于AE和Transformer的运动想象脑电信号分类研究[J]. 物联网学报, 2023, 7(1): 118-128.
[3]	张彪, 汪西明, 徐逸凡, 李文, 韩昊, 刘松仪, 陈学强. 基于多智能体深度强化学习的多域协同抗干扰方法研究[J]. 物联网学报, 2022, 6(4): 104-116.
[4]	李贤, 毕宿志, 曾泓儒, 林彬, 林晓辉. 基于智能化用户协作的边缘计算任务卸载与资源分配优化[J]. 物联网学报, 2022, 6(4): 41-52.
[5]	罗丹, 徐茹枝, 关志涛. 物联网环境中基于深度学习的差分隐私预算优化方法[J]. 物联网学报, 2022, 6(2): 65-76.
[6]	徐宣哲, 宁珂, 郑学敏, 赵明心, 徐萌萌, 吴南健, 刘力源. 基于硬件仿真系统的边缘计算人工智能视觉芯片设计验证[J]. 物联网学报, 2022, 6(1): 20-28.
[7]	李国权, 徐永海, 林金朝, 黄正文. 基于Res-DNN的端到端MIMO系统信号检测算法[J]. 物联网学报, 2022, 6(1): 65-72.
[8]	江昊, 陈宏铭, 曹以龙, 崔昊杨. 基于MIMO的大容量LPWAN技术TurMass^TM与LoRa对比[J]. 物联网学报, 2021, 5(4): 54-61.
[9]	谈玲, 荣杉山, 夏景明, SajibSarker, 马雯杰. 基于IR-VGG的多分类皮肤病实时诊断[J]. 物联网学报, 2021, 5(3): 115-125.
[10]	林椿珉, 曾烈康, 陈旭. 边缘智能驱动的高能效无人机自主导航算法研究[J]. 物联网学报, 2021, 5(2): 87-96.
[11]	廖勇,姚海梅,花远肖. 一种基于深度学习的物联网信道状态信息获取算法[J]. 物联网学报, 2019, 3(1): 8-13.