基于神经网络的OFDM信道补偿与信号检测

刘步花; 丁丹; 杨柳

doi:10.13700/j.bh.1001-5965.2019.0456

基于神经网络的OFDM信道补偿与信号检测

doi: 10.13700/j.bh.1001-5965.2019.0456

刘步花¹,
丁丹^2, ,,
杨柳²

1.
航天工程大学研究生院, 北京 101416
2.
航天工程大学电子与光学工程系, 北京 101416

基金项目:

国家“863”计划 2015AA7026085

详细信息

作者简介:
刘步花女, 硕士研究生。主要研究方向:深度学习在无线传输物理层的运用

丁丹男, 博士, 副研究员。主要研究方向:航天测控通信

通讯作者:
丁丹, E-mail: ddnjr@163.com

中图分类号: TN914.34
计量
- 文章访问数: 680
- HTML全文浏览量: 92
- PDF下载量: 136
- 被引次数: 0
出版历程
- 收稿日期: 2019-08-26
- 录用日期: 2020-02-21
- 网络出版日期: 2020-07-20

Channel compensation and signal detection of OFDM based on neural network

LIU Buhua¹,
DING Dan^{2
, ,},
YANG Liu²

1.
Department of Graduate Management, Space Engineering University, Beijing 101416, China
2.
Department of Electronic and Optical Engineering, Space Engineering University, Beijing 101416, China

Funds:

National High-tech Research and Development Program of China 2015AA7026085

More Information

Corresponding author: DING Dan, E-mail:ddnjr@163.com

摘要

摘要:
针对非线性失真和多径效应混合的复杂信道条件，提出一种基于神经网络的正交频分复用（OFDM）信道补偿与信号检测的方法。首先接收端信号利用最小二乘（LS）算法和迫零（ZF）算法做预处理，然后再输入到一层全链接层的神经网络进行进一步的信道补偿与信号检测，并恢复数据流。仿真结果表明，在没有进行输入信号功率回退（IBO）时，所提方法的误比特率（BER）性能比LS算法提升2个数量级，比线性最小均方误差（LMMSE）、最小均方误差（MMSE）提升一个数量级；在进行IBO后，所提方法能避免LS信道估计下至少4 dB的功率损失，能避免LMMSE、MMSE信道估计下至少2 dB的功率损失。所提方法在一定程度上验证了机器学习结合通信的先验知识的这种新的网络结构更能提升系统数据传输的准确率。
- 神经网络 /
- 正交频分复用(OFDM) /
- 非线性失真 /
- 多径效应 /
- 信道均衡
Abstract:
A method for Orthogonal Frequency Division Multiplexing (OFDM) channel compensation and signal detection based on neural network is proposed for the complex channel conditions of nonlinear distortion and multi-path effects. First, the receiver uses the Least Squares (LS) and Zero Forcing (ZF) algorithm to preprocess the data, and then the processed data are input to neural network with only one fully connected layer for further channel compensation and signal detection, and finally the data flow is recovered. The simulation results show that, without Input Back-Off (IBO), the Bit Error Rate (BER) performance of the proposed method is two orders of magnitude higher than that of LS algorithm, and one order of magnitude higher than that of Linear Minimum Mean Square Error (LMMSE) and Minimum Mean Square Error (MMSE); with IBO, the proposed method can avoid at least 4 dB power loss under LS channel estimation and at least 2 dB power loss under LMMSE and MMSE channel estimation. To some extent, this paper verifies that the new network structure of machine learning combined with prior knowledge of communication can improve the accuracy of data transmission.
- neural network /
- Orthogonal Frequency Division Multiplexing (OFDM) /
- nonlinear distortion /
- multipath effect /
- channel equalization

HTML全文

图 1 非线性和多径影响下的OFDM系统

Figure 1. OFDM system under influence of nonlinearity and multi-path

下载: 全尺寸图片幻灯片

图 2 神经网络接收机结构

Figure 2. Neural network receiver architecture

下载: 全尺寸图片幻灯片

图 3 发送端OFDM帧结构

Figure 3. OFDM frame structure at transmitter

下载: 全尺寸图片幻灯片

图 4 数据产生流程

Figure 4. Flowchart of data generation

下载: 全尺寸图片幻灯片

图 5 AM-AM和AM-PM非线性放大

Figure 5. AM-AM and AM-PM nonlinear amplification

下载: 全尺寸图片幻灯片

图 6 LSZF-Net与无非线性补偿的传统方法比较

Figure 6. Comparison between LSZF-Net and traditional methods without non-linear compensation

下载: 全尺寸图片幻灯片

图 7 LSZF-Net与进行IBO的传统方法比较

Figure 7. Comparison between LSZF-Net and traditional method after IBO

下载: 全尺寸图片幻灯片

图 8 不同时延扩展的多径信道LSZF-Net性能对比

Figure 8. Performance comparison of LSZF-Net in different delay spread multi-path channel

下载: 全尺寸图片幻灯片

图 9 不同网络结构性能对比

Figure 9. Performance comparison of different network structures

下载: 全尺寸图片幻灯片

表 1 训练参数设置

Table 1. Training parameter setting

OFDM参数	数值
CP长度	16
导频长度	64
信噪比/dB	0:5:25
epoch	3 000
初始学习率	0.001

注：OFDM帧结构为导频+数据；调制方式为QPSK; LSZF-Net无隐藏层；激活函数为tanh；优化器为rmsprop；损失函数为L2。

下载: 导出CSV

表 2 多径信道条件

Table 2. Multi-path channel conditions

τ_rms	信道条件
0.3T_s	h=0.544 6+1.097 5i, 0.102 9+0.207 3i, 0.019 4+0.039 2i, 0.003 7+0.007 4i
0.5T_s	h=0.821 8+0.542 5i, 0.302 3+0.199 6i, 0.111 2+0.073 4i, 0.040 9+0.027 0i, 0.015 1+0.009 9i, 0.005 5+0.003 7i
0.7T_s	h=0.525 6-0.577 8i, 0.257 3-0.282 9i, 0.126 0-0.138 5i, 0.061 7-0.067 8i, 0.030 2+0.033 2i, 0.014 8-0.016 2i, 0.007 2-008 0i, 0.003 5-0.003 9i

下载: 导出CSV

表 3 不同方法计算复杂度比较

Table 3. Comparison of computational complexity among different methods

方法	所需乘积的次数
LS	2N
MMSE	＞2N²
LMMSE	2N²
LSZF-Net	(4N)²

下载: 导出CSV

参考文献(22)

[1]	FANG X, XU Y C, CHEN Z Y, et al.Time-domain least square channel estimation for polarization-division-multiplexed CO-OFDM/OQAM systems[J].Journal of Lightwave Technology, 2015, 34(3):891-900. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=3fe63b550255f2acf5fb13e3d3a84d23
[2]	WANG J, WEN O Y, LI S Q.Soft-output MMSE MIMO detector under ML channel estimation and channel correlation[J].IEEE Signal Processing Letters, 2009, 16(8):667-670. http://d.old.wanfangdata.com.cn/NSTLQK/NSTL_QKJJ0224456164/
[3]	NOH M, LEE Y, PARK H.Low complexity LMMSE channel estimation for OFDM[J].IEE Proceedings-Communications, 2006, 153(5):645-650. http://d.old.wanfangdata.com.cn/OAPaper/oai_doaj-articles_b8296b816653c6bab7357a667adf73d2
[4]	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. http://cn.bing.com/academic/profile?id=c5b0bb6adf4ec51f581a49298f347364&encoded=0&v=paper_preview&mkt=zh-cn
[5]	NEUMANN D, WIESE T, UTSCHICK W.Learning the MMSE channel estimator[J].IEEE Transactions on Signal Processing, 2018, 66(11):2905-2917. http://d.old.wanfangdata.com.cn/OAPaper/oai_doaj-articles_79c087c0160625f7e8cf4c705211559d
[6]	SAMUEL N, DISKIN T, WIESEL A.Deep MIMO detection[C]//IEEE International Workshop on Signal Processing Advances in Wireless Communications.Piscataway: IEEE Press, 2017: 1-5.
[7]	HE H T, WEN C K, JIN S, et al.A model-driven deep learning network for MIMO detection[C]//The 6th IEEE Global Conference on Signal and Information Processing.Piscataway, IEEE Press, 2018: 584-588.
[8]	NACHMANI E, BEERY Y, BURSHTEIN D.Learning to decode linear codes using deep learning[C]//2016 54th Annual Allerton Conference on Communication, Control, and Computing.Piscataway: IEEE Press, 2016: 341-346.
[9]	NACHMANI E, MARCIANO E, LUGOSCH L, et al.Deep learning methods for improved decoding of linear codes[J].IEEE Journal of Selected Topics in Signal Processing, 2018, 12(1):119-131. http://cn.bing.com/academic/profile?id=5b6dcbdcc75d7b5ec8dc1387e794ed76&encoded=0&v=paper_preview&mkt=zh-cn
[10]	GRUBER T, CAMMERER S, HOYDIS J, et al. On deep learning-based channel decoding[C]//201751st Annual Conference on Information Sciences and Systems(CISS).Piscataway: IEEE Press, 2017: 1-6.
[11]	FEHSKE A, GAEDDERT J, REED J.A new approach to signal classification using spectral correlation and neural networks[C]//1st IEEE International Symposium on New Frontiers in Dynamic Spectrum Access Networks.Piscataway: IEEE Press, 2005: 144-150.
[12]	PENG S, JIANG H, WANG H, et al.Modulation classification based on signal constellation diagrams and deep learning[J].IEEE Transactions on Neural Networks and Learning Systems, 2019, 30(3):718-727. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=e1b8781805b3f30b68d268fa746c540b
[13]	O-SHEA T J, CORGAN J, CLANCY T C. Convolutional radio modulation recognition networks[C]//International Conference on Engineering Applications of Neural Networks. Berlin: Springer, 2016: 213-226.
[14]	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, 2019, 31(3):348-353. http://www.cnki.com.cn/Article/CJFDTotal-CASH201903009.htm LIAO Y, HUA Y X, YAO H M.Channel estimation based on deep learning for OFDM systems[J].Journal of Chongqing University of Posts and Telecommunications, 2019, 31(3):348-353(in Chinese). http://www.cnki.com.cn/Article/CJFDTotal-CASH201903009.htm
[15]	SPENCER Q H, SWINDLEHURST A L, HAARDT M.Zero-forcing methods for downlink spatial multiplexing in multiuser MIMO channels[J].IEEE Transactions on Signal Processing, 2004, 52(2):461-471. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=6aa85f959e68b155171405ff4dd6f7d7
[16]	SALEH A A M.Frequency-independent and frequency-dependent nonlinear models of TWT amplifiers[J].IEEE Transactions on Communications, 1981, 29(11):1715-1720. http://cn.bing.com/academic/profile?id=fceed0cc670e4f4e790c7865c837425f&encoded=0&v=paper_preview&mkt=zh-cn
[17]	WANG T, WEN C, WANG H, et al.Deep learning for wireless physical layer:Opportunities and challenges[J].China Communications, 2017, 14(11):92-111. http://ieeexplore.ieee.org/document/8233654
[18]	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(8):46-55. http://d.old.wanfangdata.com.cn/Periodical/dxkx201808005 ZHANG J, JIN S, WEN C K, et al.An overview of wireless transmission technology utilizing artificial intelligence[J].Telecommunications Science, 2018(8):46-55(in Chinese). http://d.old.wanfangdata.com.cn/Periodical/dxkx201808005
[19]	CHO Y S, KIM J, YANG W Y.MIMO-OFDM wireless communication with MATLAB[M].Hoboken:John Wiley & Sons, 2010:25-35.
[20]	JIANG P W, WANG T Q, HAN B, et al. Artificial intelligence-aided OFDM receiver: Design and experimental results[EB/OL].(2018-12-17)[2019-08-11].https: //arxiv.org/abs/1812.06638.
[21]	HAN D S, HWANG T.An adaptive pre-distorter for the compensation of HPA nonlinearity[J].IEEE Transactions on Broadcasting, 2000, 6(2):152-157. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=336949ad984b82b605d2779677ca6cb7
[22]	杨霖, 宋坤.OFDM系统中基于压缩感知恢复由限幅和HPA产生的非线性失真研究[J].电子学报, 2018, 46(5):1078-1083. http://d.old.wanfangdata.com.cn/Periodical/dianzixb201805008 YANG L, SONG K.Research on recovery of clipping and HPA nonlinear distortion based on compressive sensing in OFDM systems[J]. Acta Electronica Sinica, 2018, 46(5):1078-1083(in Chinese). http://d.old.wanfangdata.com.cn/Periodical/dianzixb201805008