基于调制卷积神经网络的空地数据链信道估计

刘春辉; 王美琳; 董赞亮; 王沛

doi:10.13700/j.bh.1001-5965.2020.0591

基于调制卷积神经网络的空地数据链信道估计

doi: 10.13700/j.bh.1001-5965.2020.0591

1.
北京航空航天大学无人系统研究院, 北京 100083
2.
北京航空航天大学电子信息工程学院, 北京 100083

基金项目:

科技创新2030-“新一代人工智能”重大项目 2020AAA0108200

北京市自然科学基金 4204102

详细信息

通讯作者:
刘春辉, E-mail: liuchunhui2134@buaa.edu.cn

中图分类号: V243.5; TN911.72
计量
- 文章访问数: 322
- HTML全文浏览量: 72
- PDF下载量: 40
- 被引次数: 0
出版历程
- 收稿日期: 2020-10-19
- 录用日期: 2021-01-17
- 网络出版日期: 2022-03-20

Channel estimation of air-ground data link based on modulated convolutional neural network

1.
Institute of Unmanned System, Beihang University, Beijing 100083, China
2.
School of Electronics and Information Engineering, Beihang University, Beijing 100083, China

Funds:

Science and Technology Innovation 2030-Key Project of "New Generation Artificial Intelligence" 2020AAA0108200

Beijing Municipal Natural Science Foundation 4204102

More Information

Corresponding author: LIU Chunhui, E-mail: liuchunhui2134@buaa.edu.cn

摘要

摘要:
针对复杂环境下空地数据链正交频分复用(OFDM)系统信道估计精度不足的问题，提出了一种基于调制卷积神经网络(MCNN)和双向长短时记忆网络(BiLSTM)结合的信道估计算法。利用最小二乘算法(LS)提取初始信道状态信息(CSI)；利用MCNN网络提取初始CSI的深度特征，并对网络模型进行压缩；利用BiLSTM网络对最终CSI进行预测，实现信道估计。利用构建的空地信道模型生成信道系数数据集，实现神经网络模型的训练与测试。仿真结果表明：与传统算法和现有深度学习方法相比，所提出的信道估计算法具有更小的估计误差，高信噪比条件下的系统误码率(BER)性能提升接近一个数量级；由于引入了调制滤波器技术，随着神经网络层数增加，网络模型参数量大幅减少。
- 正交频分复用(OFDM) /
- 深度学习 /
- 信道估计 /
- 空地信道模型 /
- 多径效应
Abstract:
Aimed at the inaccuracy of channel estimation of orthogonal frequency division multiplexing (OFDM) system in the complex air-ground data link environment, this paper proposes a channel estimation algorithm based on the modulated convolutional neural network (MCNN) and bidirectional long short-term memory (BiLSTM) network. First, least square (LS) algorithm is used to extract the initial channel state information (CSI), then MCNN network is used to extract the depth characteristics of the initial CSI while compressing the network model, and finally BiLSTM network is used to predict the final CSI and realize channel estimation. In the aspect of experimental verification, the air-ground channel model constructed is used to generate the channel coefficient dataset, so as to realize the training and testing of neural network model. The simulation results show that compared with the traditional methods and the existing deep learning method, the proposed channel estimation method has a lower estimation error, and the performance of the bit error ratio (BER) of the system under the condition of high SNR is improved by nearly an order of magnitude. Due to the introduction of the modulation filter technology, the number of network model parameters decreases remarkably with the increase of the number of neural network layers.
- orthogonal frequency division multiplexing (OFDM) /
- deep learning /
- channel estimation /
- air-ground channel model /
- multipath effect

HTML全文

图 1 基于OFDM的通信系统结构

Figure 1. Communication system structure based on OFDM

下载: 全尺寸图片幻灯片

图 2 二径空地信道模型

Figure 2. Two-ray air-ground channel model

下载: 全尺寸图片幻灯片

图 3 基于深度学习的信道估计框架

Figure 3. Channel estimation framework based on deep learning

下载: 全尺寸图片幻灯片

图 4 基于调制滤波器的调制过程

Figure 4. Modulation process based on modulation filter

下载: 全尺寸图片幻灯片

图 5 BiLSMT网络结构

Figure 5. BiLSTM network structure

下载: 全尺寸图片幻灯片

图 6 MC-BI网络模型的训练与应用

Figure 6. Training and application of MC-BI network model

下载: 全尺寸图片幻灯片

图 7 三种典型场景下不同信道估计算法的NMSE比较

Figure 7. NMSE comparison of different channel estimation algorithms in three typical scenes

下载: 全尺寸图片幻灯片

图 8 三种典型场景下不同信道估计算法的BER曲线

Figure 8. BER curves of different channel estimation algorithms in three typical scenes

下载: 全尺寸图片幻灯片

图 9 压缩前后模型参数量比较

Figure 9. Comparison of model parameters before and after compression

下载: 全尺寸图片幻灯片

图 10 MCNN+BiLSTM与CNN+BiLSTM网络BER对比

Figure 10. BER comparison between MCNN+BiLSTM network and CNN+BiLSTM network

下载: 全尺寸图片幻灯片

图 11 不同BiLSTM层与不同MCNN层结合的MC-BI算法的BER曲线

Figure 11. BER curves of MC-BI algorithm with different BiLSTM layers and MCNN layers

下载: 全尺寸图片幻灯片

图 12 三种MC-BI网络的BER比较

Figure 12. BER comparison of three MC-BI networks

下载: 全尺寸图片幻灯片

表 1 信道模型参数

Table 1. Channel model parameters

场景	h_A/m	h_G/m	d/km	n₁	n₂	s_g/m
郊区	1 000	20	46	15	1	0.05
山区	4 000	20	60	5	1	1 000
海上	800	20	25	81	1	0.187 2

下载: 导出CSV

表 2 OFDM系统参数设置

Table 2. OFDM system parameter setup

参数	数值
子载波数	64
OFDM符号数	64
CP长度	8
导频间隔	8
调制方式	4QAM

下载: 导出CSV

参考文献(26)

[1]	丛伟, 李宏. 基于OFDM的无人机通信链路[C]//2006中国无人机大会, 2006: 629-634. CONG W, LI H. Communication link of UAV based on OFDM[C]//2006 China UAV Conference, 2006: 629-634(in Chinese).
[2]	章勇. OFDM系统信道估计技术研究[D]. 北京: 北京邮电大学, 2008: 10-12. ZHANG Y. Research on channel estimation techniques of OFDM system[D]. Beijing: Beijing University of Posts and Telecommunications, 2008: 10-12(in Chinese).
[3]	董秀洁, 王莉, 王素菊. 一种改进的LS信道估计算法[J]. 哈尔滨理工大学学报, 2009, 14(1): 47-50. doi: 10.3969/j.issn.1007-2683.2009.01.012 DONG X J, WANG L, WANG S J. An improved LS channel estimation algorithm[J]. Journal of Harbin University of Science and Technology, 2009, 14(1): 47-50(in Chinese). doi: 10.3969/j.issn.1007-2683.2009.01.012
[4]	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. doi: 10.1109/LSP.2009.2021368
[5]	ZHANG K, XUE L S, LIU X P, et al. Performance comparison of LS, LMMSE channel estimation method in frequency and time domain for OQAM/OFDM systems[C]//2017 IEEE 17th International Conference on Communication Technology (ICCT). Piscataway: IEEE Press, 2017: 224-228.
[6]	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, 2018, 7(1): 114-117. doi: 10.1109/LWC.2017.2757490
[7]	HE H T, WEN C K, JIN S, et al. A model-driven deep learning network for MIMO detection[C]//2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP). Piscataway: IEEE Press, 2018: 584-588.
[8]	GAO X X, JIN S, WEN C K, et al. ComNet: Combination of deep learning and expert knowledge in OFDM receivers[J]. IEEE Communications Letters, 2018, 22(12): 2627-2630. doi: 10.1109/LCOMM.2018.2877965
[9]	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(自然科学版), 2019, 31(3): 348-353. https://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 (Natural Science Edition), 2019, 31(3): 348-353(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-CASH201903009.htm
[10]	廖勇, 花远肖, 姚海梅, 等. 高速移动环境下基于深度学习的信道估计方法[J]. 电子学报, 2019, 47(8): 1701-1707. doi: 10.3969/j.issn.0372-2112.2019.08.013 LIAO Y, HUA Y X, YAO H M, et al. Channel estimation method based on deep learning in high-speed mobile environments[J]. Acta Electronica Sinica, 2019, 47(8): 1701-1707(in Chinese). doi: 10.3969/j.issn.0372-2112.2019.08.013
[11]	SAINATH T N, VINYALS O, SENIOR A, et al. Convolutional, long short-term memory, fully connected deep neural networks[C]//2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). Piscataway: IEEE Press, 2015: 4580-4584.
[12]	LIAO Y, HUA Y X, CAI Y L. Deep learning based channel estimation algorithm for fast time-varying MIMO-OFDM systems[J]. IEEE Communications Letters, 2020, 24(3): 572-576. doi: 10.1109/LCOMM.2019.2960242
[13]	刘步花, 丁丹, 杨柳. 基于神经网络的OFDM信道补偿与信号检测[J]. 北京航空航天大学学报, 2020, 46(7): 1363-1370. doi: 10.13700/j.bh.1001-5965.2019.0456 LIU B H, DING D, YANG L. Channel compensation and signal detection of OFDM based on neural network[J]. Journal of Beijing University of Aeronautics and Astronautics, 2020, 46(7): 1363-1370(in Chinese). doi: 10.13700/j.bh.1001-5965.2019.0456
[14]	ZHANG J, WEN C K, JIN S, et al. Artificial intelligence-aided receiver for a CP-free OFDM system: Design, simulation, and experimental test[J]. IEEE Access, 2019, 7: 58901-58914. doi: 10.1109/ACCESS.2019.2914928
[15]	MATOLAK D W, SUN R Y. Air-ground channel characterization for unmanned aircraft systems—Part Ⅰ: Methods, measurements, and models for over-water settings[J]. IEEE Transactions on Vehicular Technology, 2017, 66(1): 26-44. doi: 10.1109/TVT.2016.2530306
[16]	MATOLAK D W, SUN R Y. Air-ground channel characterization for unmanned aircraft systems—Part Ⅱ: Hilly and mountainous settings[J]. IEEE Transactions on Vehicular Technology, 2017, 66(3): 1913-1925. doi: 10.1109/TVT.2016.2585504
[17]	MATOLAK D W, SUN R Y. Air-ground channel characterization for unmanned aircraft systems—Part Ⅱ: The suburban and nearurban environments[J]. IEEE Transactions on Vehicular Technology, 2017, 66(8): 6607-6618. doi: 10.1109/TVT.2017.2659651
[18]	HAN S, POOL J, TRAN J, et al. Learning both weights and connections for efficient neural networks[EB/OL]. (2015-10-30)[2020-10-15]. https://arxiv.org/abs/1506.02626v3.
[19]	HAN S, MAO H Z, DALLY W J. Deep compression: Compressing deep neural networks with pruning, trained quantization and huffman coding[EB/OL]. (2016-02-15)[2020-10-15]. https://arxiv.org/abs/1510.00149.
[20]	SZEGEDY C, VANHOUCKE V, IOFFE S, et al. Rethinking the inception architecture for computer vision[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE Press, 2016: 2818-2826.
[21]	WANG X D, ZHANG B C, LI C, et al. Modulated convolutional networks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2018: 840-848.
[22]	SCHUSTER M, PALIWAL K K. Bidirectional recurrent neural networks[J]. IEEE Transactions on Signal Processing, 1997, 45(11): 2673-2681. doi: 10.1109/78.650093
[23]	姚宜东, 徐毅, 杨新华, 等. 介质表面电磁波反射和透射能量的分析与计算[J]. 通信技术, 2016, 49(5): 558-562. doi: 10.3969/j.issn.1002-0802.2016.05.010 YAO Y D, XU Y, YANG X H, et al. Analysis and calculation on reflection and transmission energy of electromagnetic wave at dielectric interface[J]. Communications Technology, 2016, 49(5): 558-562(in Chinese). doi: 10.3969/j.issn.1002-0802.2016.05.010
[24]	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. doi: 10.1109/LWC.2018.2818160
[25]	ZHANG D N, DING W R, LIU C H, et al. Modulated autocorrelation convolution networks for automatic modulation classification based on small sample set[J]. IEEE Access, 2020, 8: 27097-27105. doi: 10.1109/ACCESS.2020.2971586
[26]	DAO D N, TELLAMBURA C. Intercarrier interference self-cancellation space-frequency codes for MIMO-OFDM[J]. IEEE Transactions on Vehicular Technology, 2005, 54(5): 1729-1738. doi: 10.1109/TVT.2005.853477