注意力机制CNN的毫米波大规模MIMO系统信道估计算法

doi:10.12305/j.issn.1001-506X.2022.01.38

摘要/Abstract

摘要：

在室外光线追踪通信场景下, 针对毫米波大规模多输入多输出(multiple input multiple output, MIMO)信道具有稀疏特性、系统受噪声因素影响导致信道估计精度低的问题, 提出一种基于图像去噪的注意力机制卷积神经网络信道估计方法。首先, 设定参数产生模拟真实环境的数据集, 将所产生的信道矩阵看作二维图像。然后, 构建注意力机制网络以增强图像中噪声特征的显著性, 并将注意力机制网络嵌入卷积神经网络中进行特征融合。最后, 通过网络模型提取噪声并得到去噪的图像, 即估计信道矩阵。仿真结果表明, 与最小二乘法(least square, LS)、最小均方误差(minimum mean square error, MMSE)、卷积神经网络(convolutional neural network, CNN)和去噪CNN (denoising CNN, DnCNN)算法相比, 所提出的Attention-CNN方法信道估计精度平均提升约1.86 dB。

关键词: 毫米波大规模多输入多输出, 信道估计, 卷积神经网络, 注意力机制, 去噪

Abstract:

In outdoor ray tracing communication scenarios, aiming at the problems of low channel estimation accuracy defected by its sparse characteristics and noise factors in millimeter-wave massive multiple input multiple output (MIMO), an image denoising attention mechanism based convolutional neural network channel estimation algorithm is proposed. Firstly, after constructing the data set for simulating the real environment by setting the parameters, generate channel matrix is regarded as a two-dimensional image. Then, the attention mechanism network is constructed to enhance the saliency of the noise features in the image, and the attention mechanism network is embedded in the convolutional neural network (CNN) for feature fusion. Finally, the noise extracted by the network model achieves the denoising effect and the denoised image, the estimated channel matrix is obtained. The simulation results demonstrate that the proposed attention mechanism based CNN (Attention-CNN) algorithm achieves better performance of higher channel estimation accuracy, which improved by about 1.86 dB on average, compared with least square (LS), minimum mean square error (MMSE), CNN and denoising convolutional neural network (DnCNN).

Key words: millimeter-wave massive multiple input multiple output (MIMO), channel estimation, convolutional neural network (CNN), attention mechanism, denoising

中图分类号:

TP391.9

刘紫燕, 马珊珊, 梁静, 朱明成, 袁磊. 注意力机制CNN的毫米波大规模MIMO系统信道估计算法[J]. 系统工程与电子技术, 2022, 44(1): 307-312.

Ziyan LIU, Shanshan MA, Jing LIANG, Mingcheng ZHU, Lei YUAN. Attention mechanism based CNN channel estimation algorithm in millimeter-wave massive MIMO system[J]. Systems Engineering and Electronics, 2022, 44(1): 307-312.

图/表 10

图1

图2

图3

表1

表2

图4

图5

图6

图7

图8

参考文献 30

1	MUMTAZ S , RODRIQUEZ J , DAI L L . mmWave massive MIMO: a paradigm for 5G[M]. New York: Academic, 2016.
2	BOCCARDI F , HEATH R , LOZANO A , et al. Five disruptive technology directions for 5G[J]. IEEE Communications Magazine, 2014, 52 (2): 74- 80. doi: 10.1109/MCOM.2014.6736746
3	康国良. 毫米波大规模MIMO系统信道估计算法研究[D]. 南昌: 江西理工大学, 2018.
	KANG G L. Research on channel estimation algorithm for millimeter wave massive MIMO system[D]. Nanchang: Jiangxi University of Science and Technology, 2018.
4	SWINDLEHURST A L , AYANOGLU E , HEYDARI P , et al. Millimeter-wave massive MIMO: the next wireless revolution?[J]. IEEE Communication Magazine, 2014, 52 (9): 56- 62. doi: 10.1109/MCOM.2014.6894453
5	SUN Z H, ZHAO Z, FU X M, et al. Limited feedback double directional massive MIMO channel estimation: from low-rank modeling to deep learning[C]//Proc. of the IEEE 19th International Workshop on Signal Processing Advances in Wireless Communications, 2018.
6	HUANG X, LIU S M. Massive MIMO channel estimation for vehicular communications: a deep learning based approach[C]//Proc. of the IEEE International Conference on Communications Workshops, 2020.
7	DEMIR O T , BJORNSON E . Channel estimation in massive MIMO under hardware non-linearities: Bayesian methods versus deep learning[J]. IEEE Open Journal of the Communications Society, 2020, 32 (8): 109- 124.
8	YE H , LI G Y , JUAN G B . Power of deep learning for channel estimation and signal detection in OFDM systems[J]. IEEE Wireless Communications Letters, 2018, 7 (1): 114- 117. doi: 10.1109/LWC.2017.2757490
9	ANG H J , SONG Y , YANG J H , et al. Deep-learning-based millimeter-wave massive MIMO for hybrid precoding[J]. IEEE Trans.on Vehicular Technology, 2019, 68 (3): 3027- 3032. doi: 10.1109/TVT.2019.2893928
10	MA J J , PING L . Data-aided channel estimation in large antenna systems[J]. IEEE Trans.on Signal Processing, 2014, 62 (12): 3111- 3124. doi: 10.1109/TSP.2014.2321120
11	CHEN X Y , JIANG M F . Adaptive statistical Bayesian MMSE channel estimation for visible light communication[J]. IEEE Trans.on Signal Processing, 2017, 65 (5): 1287- 1299. doi: 10.1109/TSP.2016.2630036
12	GAO Z , DAI L L , WANG Z , et al. Spatially common sparsity based adaptive channel estimation and feedback for FDD massive MIMO[J]. IEEE Trans.on Signal Processing, 2015, 63 (23): 6169- 6183. doi: 10.1109/TSP.2015.2463260
13	邵凯, 陈连成, 刘胤. 高移动性Jakes信道的学习与估计[J]. 系统工程与电子技术, 2020, 43 (4): 1119- 1125.
	SHAO K , CHEN L C , LIU Y . Learning and estimation of high mobility Jakes channel[J]. Systems Engineering and Electro-nics, 2020, 43 (4): 1119- 1125.
14	LUO C Z , JI J , WANG Q Y , et al. Channel state information prediction for 5G wireless communications: a deep learning approach[J]. IEEE Trans.on Network Science and Engineering, 2020, 7 (1): 227- 236. doi: 10.1109/TNSE.2018.2848960
15	SABETI P, FARHANG A, MACALUSO I, et al. Blind channel estimation for massive MIMO: a deep learning assisted approach[C]//Proc. of the IEEE International Conference on Communications, 2020.
16	GAO Z F , WANG Y , LIU X F , et al. FFDNet-based channel estimation for massive MIMO visible light communication systems[J]. IEEE Wireless Communications Letters, 2020, 9 (3): 340- 343. doi: 10.1109/LWC.2019.2954511
17	GU J W, SHAN C, CHEN X L, et al. A novel pilot-aided channel estimation scheme based on RNN for FDD-LTE systems[C]//Proc. of the 10th International Conference on Wireless Communications and Signal Processing, 2018.
18	JIN J , ZHANG J M , JIN S W , et al. Channel estimation for cell-free mmWave massive MIMO through deep learning[J]. IEEE Trans.on Vehicular Technology, 2019, 68 (10): 10325- 10329. doi: 10.1109/TVT.2019.2937543
19	ZHANG K , ZUO W L , CHEN Y Q , et al. Beyond a Gaussian denoiser: residual learning of deep CNN for image denoising[J]. IEEE Trans.on Image Processing, 2017, 26 (7): 3142- 3155. doi: 10.1109/TIP.2017.2662206
20	罗仁泽, 王瑞杰, 张可, 等. 残差卷积自编码网络图像去噪方法[J]. 计算机仿真, 2021, 38 (5): 455- 461. doi: 10.3969/j.issn.1006-9348.2021.05.093
	LUO R Z , WANG R J , ZHANG K , et al. Residual convolution self-encoding network image denoising method[J]. Computer Simulation, 2021, 38 (5): 455- 461. doi: 10.3969/j.issn.1006-9348.2021.05.093
21	杜渺勇, 于祥雨, 周浩, 等. 基于自编码卷积神经网络的图像去噪算法[J]. 杭州师范大学学报(自然科学版), 2021, 20 (1): 95- 101.
	DU M Y , YU X Y , ZHOU H , et al. Image denoising algorithm based on self-encoding convolutional neural network[J]. Journal of Hangzhou Normal University (Natural Science Edition), 2021, 20 (1): 95- 101.
22	卢文侣. 毫米波大规模MIMO系统信道估计研究[D]. 北京: 北京邮电大学, 2017.
	LU W L. Channel estimation study of millimeter wave large-scale MIMO system[D]. Beijing: Beijing University of Posts and Telecommunications, 2017.
23	HE H , WEN C K , JIN S , et al. Deep learning-based channel estimation for beamspace mmWave massive MIMO systems[J]. IEEE Wireless Communication Letters, 2018, 7 (5): 852- 855. doi: 10.1109/LWC.2018.2832128
24	LIN Z, FENG M, SANTOS C N, et al. A structured self-attentive sentence embedding[EB/OL]. [2021-06-22]. ar Xiv preprint ar Xiv: 1703.03130, 2017.
25	ZHANG K, ZUO W Y, GU S, et al. Learning deep CNN denoiser prior for image restoration[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 3929-3938.
26	WANG J K , ZHENG Y , WANG M H , et al. Object-scale adaptive convolutional neural networks for high-spatial resolution remote sensing image classification[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2021, 14, 283- 299. doi: 10.1109/JSTARS.2020.3041859
27	TAMURA H, YANGAGISAWA K, SHIRANE A, et al. Wireless devices identification with light-weight convolutional neural network operating on quadrant IQ transition image[C]//Proc. of the 18th IEEE International New Circuits and Systems Conference, 2020: 106-109.
28	GUO S, YAN Z, ZHANG K P, et al. Toward convolutional blind denoising of real photographs[C]//Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019: 1712-1722.
29	周飞燕, 金林鹏, 董军. 卷积神经网络研究综述[J]. 计算机学报, 2017, 40 (6): 1229- 1251.
	ZHOU F Y , JIN L P , DONG J . A review of research on convolutional neural networks[J]. Acta computerica sinica, 2017, 40 (6): 1229- 1251.
30	ALKHATEEB A. DeepMIMO: a generic deep learning dataset for millimeter wave and massive MIMO applications[C]//Proc. of the Information Theory and Applications Workshop, 2019.

算法	乘法运算次数	复杂度
LS	U²+N	O(U²)
MMSE	4U³+N	O(U³)
CNN	102U³B+36L	O(U³B)
DnCNN	128U²B+24L	O(U²B)
Attention-CNN	176U²B+32L	O(U²B)

参数	数值
基站数	5, 6, 7, 8
用户数	1 200~1 800
基站天线数	M_x=1, M_y=16, M_z=8
天线间距	0.5
系统带宽/GHz	0.5
OFDM子载波数	1 024
路径数	5
学习率γ	0.001
激活函数	Relu
SNR范围/dB	30
bach_size	32
epoch	300

[1]	宋爽, 张悦, 张琳娜, 岑翼刚, 李浥东. 基于深度学习的轻量化目标检测算法[J]. 系统工程与电子技术, 2022, 44(9): 2716-2725.
[2]	韩啸, 陈世文, 陈蒙, 杨锦程. 基于互易点学习的LPI信号开集识别[J]. 系统工程与电子技术, 2022, 44(9): 2752-2759.
[3]	王彩云, 吴钇达, 王佳宁, 马璐, 赵焕玥. 基于改进的CNN和数据增强的SAR目标识别[J]. 系统工程与电子技术, 2022, 44(8): 2483-2487.
[4]	韦娟, 杨皇卫, 宁方立. 基于NMF与CNN联合优化的声学场景分类[J]. 系统工程与电子技术, 2022, 44(5): 1433-1438.
[5]	徐平亮, 崔亚奇, 熊伟, 熊振宇, 顾祥岐. 生成式中断航迹接续关联方法[J]. 系统工程与电子技术, 2022, 44(5): 1543-1552.
[6]	陈冬, 句彦伟. 基于语义分割实现的SAR图像舰船目标检测[J]. 系统工程与电子技术, 2022, 44(4): 1195-1201.
[7]	方伟, 王玉, 闫文君, 林冲. 基于神经网络的符号化飞行动作识别[J]. 系统工程与电子技术, 2022, 44(3): 737-745.
[8]	孙晶明, 虞盛康, 孙俊. 基于深度学习的HRRP识别姿态敏感性分析[J]. 系统工程与电子技术, 2022, 44(3): 802-807.
[9]	党建, 李业伟, 朱永东, 郭荣斌, 张在琛, 吴亮. 可重构智能表面通信系统的渐进信道估计方法[J]. 系统工程与电子技术, 2022, 44(3): 998-1006.
[10]	吴灏, 康颖, 葛松虎, 李亚星, 孟进. VHF/UHF通信电台接收链路中的干扰抑制合并方法[J]. 系统工程与电子技术, 2022, 44(3): 1014-1021.
[11]	刘恒燕, 张立民, 闫文君, 钟兆根, 凌青, 梁晓军. 基于WBP-CNN算法的LDPC译码[J]. 系统工程与电子技术, 2022, 44(3): 1030-1035.
[12]	邵凯, 朱苗苗, 王光宇. 基于生成对抗与卷积神经网络的调制识别方法[J]. 系统工程与电子技术, 2022, 44(3): 1036-1043.
[13]	吴涛, 王伦文, 朱敬成. 基于迁移学习和注意力机制的伪装图像分割[J]. 系统工程与电子技术, 2022, 44(2): 376-384.
[14]	张玺, 金正猛, 姜亚琴. 融合深度图像先验的全变差图像着色算法[J]. 系统工程与电子技术, 2022, 44(2): 385-393.
[15]	金涛, 王晓峰, 田润澜, 张歆东. 基于改进1DCNN+TCN的雷达辐射源快速识别方法[J]. 系统工程与电子技术, 2022, 44(2): 463-469.