基于深度学习的正交频分复用系统信道估计

doi:10.11959/j.issn.1000-436x.2023240

摘要/Abstract

摘要：

针对5G系统信号接收子载波间串扰和子符号间干扰问题，提出了一种高效的基于深度学习的信道估计模型。在导频处进行初步估计获得估计信道，并将其视为含噪声的低分辨率图像样本输入信道估计模型，通过学习低分辨率图像与高分辨率图像之间的映射关系，最终去除输入信道的噪声，还原高分辨率信道图像，获得整个信道状态信息。仿真结果表明，该模型不仅延续了传统注意力机制抑制冗余信息的优势，降低了计算开销，还能获得良好的精度和鲁棒性，对各种信道都有较好的估计效果。

关键词: 深度学习, 信道估计, 图像恢复, 注意力机制

Abstract:

An efficient channel estimation model based on deep learning was proposed for the problems of inter-carrier interference and inter-symbol interference in 5G system signal reception.The estimated channels were obtained through a preliminary estimation at the pilots.And they were treated as low resolution images containing noise, which were input into the channel estimation model.By learning the mapping relationship between the low resolution images and the high resolution images, the noise in input channels was removed, and the high-resolution channel images were restored to obtain the entire channel state information eventually.The simulation results show that the model not only continues the advantages of traditional attention mechanisms in suppressing redundant information, reduces computational overhead, but also achieves good accuracy and robustness, and has good estimation performance for various channels.

Key words: deep learning, channel estimation, image restoration, attention mechanism

中图分类号:

TN92

张昀, 周婧, 黄经纬, 于舒娟, 黄丽亚. 基于深度学习的正交频分复用系统信道估计[J]. 通信学报, 2023, 44(12): 124-133.

Yun ZHANG, Jing ZHOU, Jingwei HUANG, Shujuan YU, Liya HUANG. Channel estimation for OFDM system based on deep learning[J]. Journal on Communications, 2023, 44(12): 124-133.

图/表 18

图1

图2

图3

图4

图5

图6

表1

表2

图7

图8

表3

表4

图9

图10

图11

图12

图13

表5

参考文献 26

[1]	任斌, 张振宇, 方荣一 ,等. 面向 5G-Advanced 无线系统的高精度定位技术[J]. 电信科学, 2022,38(3): 65-73.
	REN B , ZHANG Z Y , FANG R Y ,et al. High-accuracy positioning for 5G-Advanced wireless systems[J]. Telecommunications Science, 2022,38(3): 65-73.
[2]	何彬, 李国兵, 陈源 ,等. 基于图信号处理的 OFDM 系统导频设计和信道估计方法[J]. 物联网学报, 2022,6(3): 91-102.
	HE B , LI G B , CHEN Y ,et al. Graph signal processing based pilot pattern design and channel estimation for OFDM system[J]. Chinese Journal on Internet of Things, 2022,6(3): 91-102.
[3]	桂冠, 王禹, 黄浩 . 基于深度学习的物理层无线通信技术:机遇与挑战[J]. 通信学报, 2019,40(2): 19-23.
	GUI G , WANG Y , HUANG H . Deep learning based physical layer wireless communication techniques:opportunities and challenges[J]. Journal on Communications, 2019,40(2): 19-23.
[4]	XU W H , WU Z Z , UENG Y L ,et al. Improved polar decoder based on deep learning[C]// Proceedings of IEEE International Workshop on Signal Processing Systems (SiPS). Piscataway:IEEE Press, 2017: 1-6.
[5]	YAN L C , WANG Y , ZHENG N . 5G massive MIMO signal detection algorithm based on deep learning[J]. Computational Intelligence and Neuroscience, 2022,2022: 1-9.
[6]	CHENG X , LIU D J , WANG C ,et al. Deep learning-based channel estimation and equalization scheme for FBMC/OQAM systems[J]. IEEE Wireless Communications Letters, 2019,8(3): 881-884.
[7]	黄源, 何怡刚, 吴裕庭 ,等. 基于深度学习的压缩感知FDD大规模MIMO系统稀疏信道估计算法[J]. 通信学报, 2021,42(8): 61-69.
	HUANG Y , HE Y G , WU Y T ,et al. Deep learning for compressed sensing based sparse channel estimation in FDD massive MIMO systems[J]. Journal on Communications, 2021,42(8): 61-69.
[8]	YE H , LI G Y , JUANG B H . Power of deep learning for channel estimation and signal detection in OFDM systems[J]. IEEE Wireless Communications Letters, 2017,7(1): 114-117.
[9]	HE H T , WEN C K , JIN S ,et al. Deep learning-based channel estimation for beamspace mmWave massive MIMO systems[J]. IEEE Wireless Communications Letters, 2018,7(5): 852-855.
[10]	HUANG H J , YANG J , HUANG H ,et al. Deep learning for super-resolution channel estimation and DOA estimation based massive MIMO system[J]. IEEE Transactions on Vehicular Technology, 2018,67(9): 8549-8560.
[11]	YANG Y W , GAO F F , MA X L ,et al. Deep learning-based channel estimation for doubly selective fading channels[J]. IEEE Access, 2019,7: 36579-36589.
[12]	王莹, 任军, 史可 ,等. 基于深度学习的广义频分复用系统时频双选择信道估计[J]. 通信学报, 2021,42(10): 233-242.
	WANG Y , REN J , SHI K ,et al. Doubly-selective channel estimation for generalized frequency division multiplexing systems based on deep learning[J]. Journal on Communications, 2021,42(10): 233-242.
[13]	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.
[14]	SOLTANI M , POURAHMADI V , MIRZAEI A ,et al. Deep learning-based channel estimation[J]. IEEE Communications Letters, 2019,23(4): 652-655.
[15]	JIN Y , ZHANG J Y , AI B ,et al. Channel estimation for mmWave massive MIMO with convolutional blind denoising network[J]. IEEE Communications Letters, 2020,24(1): 95-98.
[16]	WANG Y , HUANG H , XU Q ,et al. Practical deep raw image denoising on mobile devices[C]// Proceedings of Computer Vision-ECCV 2020. Berlin:Springer, 2020: 1-16.
[17]	ZHAO J H , LIU J , YANG L H ,et al. Future 5G-oriented system for urban rail transit:opportunities and challenges[J]. China Communications, 2021,18(2): 1-12.
[18]	MUMTAZ S , RODRIGUEZ J , DAI L . mmWave massive MIMO:a paradigm for 5G[M]. Pittsburgh: Academic Press, 2016.
[19]	ALKHATEEB A , AYACH O E , LEUS G ,et al. Channel estimation and hybrid precoding for millimeter wave cellular systems[J]. IEEE Journal of Selected Topics in Signal Processing, 2014,8(5): 831-846.
[20]	WANG Y C , XU W , ZHANG H ,et al. Wideband mmWave channel estimation for hybrid massive MIMO with low-precision ADCs[J]. IEEE Wireless Communications Letters, 2019,8(1): 285-288.
[21]	DONG P , ZHANG H , LI G Y ,et al. Deep CNN-based channel estimation for mmWave massive MIMO systems[J]. IEEE Journal of Selected Topics in Signal Processing, 2019,13(5): 989-1000.
[22]	CHOLLET F . Xception:deep learning with depthwise separable convolutions[C]// Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway:IEEE Press, 2017: 1800-1807.
[23]	CHEN H , GU J , ZHANG Z . Attention in attention network for image super-resolution[J]. arXiv Preprint,arXiv:2104.09497, 2021.
[24]	CHEN Y P , DAI X Y , LIU M C ,et al. Dynamic convolution:attention over convolution kernels[C]// Proceedings of 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway:IEEE Press, 2020: 11027-11036.
[25]	LOSHCHILOV I , HUTTER F . Decoupled weight decay regularization[J]. arXiv Preprint,arXiv:1711.05101, 2017.
[26]	3GPP. Study on channel model for frequencies from 0.5 to 100 GHz (Release 16):TR 38.901 V16.1.0[S]. 2019.

参数	参数值
系统带宽/MHz	1.6
载波频率/GHz	2.1
用户设备速度/(km·h^-1)	50
子载波数	72
子符号数	14
导频数	48

模块	层数	滤波器数量	滤波器大小	激活函数
SRCNN	1	64	9×9	ReLU
	2	32	1×1	ReLU
	3	2	5×5	—
DNCNN	4	64	3×3	ReLU
	5～22	64	3×3	ReLU
	23	2	3×3	—

参数名称	参数取值
载波频率/GHz	28
采样频率/MHz	100
接收天线数	16
发射天线数	32
子载波数	64
视距角	60°
基站天线高度/m	10
主路径数	1～4
最大多普勒频移/Hz	0,500,1 000

层数	滤波器数量	滤波器大小	激活函数
1～9	64	3×3	ReLU
10	2	3×3	Tanh

算法	参数量/个	浮点运算/FLOP
ChannelNet	0.8×10⁶	3 7620 9408
DNCNN	0.7×10⁶	3 4386 7392
Deep CNN	0.04×10⁶	1 1442 5856
AAB	0.05×10⁶	2541 6100
PDAAB	1.1×10⁶	3352 3364