正交频分复用（Orthogonal Frequency Division Multiplexing，OFDM）是无线通信系统的关键技术之一，其具有良好的抗多径衰落能力和频谱利用率。然而，在无线通信过程中信道的多径效应和多普勒频移会导致信号在传输过程中失真，使得OFDM通信接收端的信息恢复变得困难。因此，针对复杂场景下OFDM传统通信系统接收端的信息恢复的问题，论文基于卷积神经网络，围绕OFDM通信接收端的信息恢复开展了研究工作。主要工作有以下两个方面：

       （1）提出了一种基于卷积神经网络（Convolutional Neural Network，CNN）的OFDM通信智能接收方法。该方法采用卷积神经网络来替代传统OFDM接收端的所有串行模块，对通信接收端进行整体优化，从而避免了传统接收端的模块化的串行误差累积和复杂的导频操作。针对OFDM通信接收端信号的处理，采用了一维卷积神经网络模型取代OFDM通信接收端。在此基础上，设计改进了DenseNet和MobileNetV3卷积神经网络结构，提出了基于DenseNet和MobileNetV3的OFDM通信智能接收方法。在不同仿真条件下实验结果表明下，与传统接收方法相比，OFDM通信智能接收方法能够降低误码率，提高OFDM通信系统的性能。

       （2）在前面OFDM通信智能接收方法的基础上进一步加入了最小二乘（Least Square，LS）信道估计先验知识，将数据驱动与知识相结合，提出了基于双通道卷积神经网络的OFDM通信智能接收方法。设计了一种双通道卷积神经网络（Dual-path Convolutional Neural Network，DCNet）作为OFDM通信智能接收方法的实现方式。该卷积神经网络通过两个不同卷积通道对数据集进行特征提取，将提取到的不同特征信息经过特征融合模块形成一个包含两种特征信息的新特征。在具有复杂度较低并且适用性较好WINNER Ⅱ信道模型下进行了部分仿真实验，在不同仿真条件下实验结果表明该方法能够进一步提高OFDM通信系统的性能并且对脉冲噪声也有较好的抑制作用。

      Orthogonal Frequency Division Multiplexing (OFDM) is one of the key technologies for wireless communication systems, which has good resistance to multipath fading and spectrum utilization. However, the multipath effect and Doppler shift of the channel in the wireless communication process can cause distortion of the signal during transmission, making it difficult to recover the information at the receiver end of OFDM communication. Therefore, to address the problem of information recovery at the receiver end of OFDM conventional communication systems in complex scenarios, the thesis conducts research work based on convolutional neural networks around information recovery at the receiver end of OFDM communication. The main work has two aspects as follows:

(1) A Convolutional Neural Network (CNN) based intelligent reception method for OFDM communication is proposed. The method uses a convolutional neural network to replace all the serial modules of the conventional OFDM receiver to optimize the communication receiver as a whole, thus avoiding the modular serial error accumulation and complex frequency-guiding operations of the conventional receiver. For the processing of the signal at the OFDM communication receiver, a one-dimensional convolutional neural network model is used to replace the OFDM communication receiver. Based on this, the DenseNet and MobileNetV3 convolutional neural network structures are designed and improved, and the intelligent reception method of OFDM communication based on DenseNet and MobileNetV3 is proposed. The experimental results under different simulation conditions show that the intelligent reception method for OFDM communication can reduce the BER and improve the performance of OFDM communication system compared with the traditional reception method.

(2) Based on the previous intelligent reception method for OFDM communication, the least-squares (LS) channel estimation a priori knowledge is further added to combine data-driven and knowledge, and an intelligent reception method for OFDM communication based on dual-path convolutional neural network is proposed. A Dual-path Convolutional Neural Network (DCNet) is designed as an implementation of the intelligent reception method for OFDM communication. The convolutional neural network is trained on the data set by two different channels, and the extracted features with different information are fused to form a new feature containing information of both features. Simulation experiments are carried out under the WINNER II channel model with low complexity and good applicability. The experimental results show that the method can further improve the performance of OFDM communication system and has a good suppression effect on impulse noise under different simulation conditions.

[1] Huang H, Miao W, Min G, et al. NFV and Blockchain Enabled 5G for Ultra-Reliable and Low-Latency Communications in Industry: Architecture and Performance Evaluation[J]. IEEE Transactions on Industrial Informatics, 2020, 17(8): 5595-5604.

[2] Agiwal M, Kwon H, Park S, et al. A Survey on 4G-5G Dual Connectivity: Road to 5G Implementation[J]. IEEE Access, 2021, 9:16193-16210.

[3] 卢忠勉. 基于软件无线电的OFDM系统智能抗干扰方法研究[D].吉林大学,2021.

[4] 卢为党, 吴宣利, 沙学军等. OFDM固定中继系统中单中继多源用户资源分配方案[J].通信学报,2011,32(09):26-32.

[5] Zhang C, Li D. Fast Fading Channel Estimation for OFDM Systems with Complexity Reduction[J]. Chinese Journal of Electronics,2021,30(06):1173-1177.

[6] 李果, 文妮, 宫丰奎等. 部分频带干扰下的OFDM系统干扰检测与分集抑制算法[J].通信学报, 2021,42(09):194-204.

[7] 李莉. MIMO-OFDM系统原理、应用及仿真[M]. 机械工业出版社, 2014.

[8] Otter D , Medina J , Kalita J . A Survey of the Usages of Deep Learning for Natural Language Processing[J]. IEEE Transactions on Neural Networks and Learning Systems, 32(2):604-624.

[9] Shaqwi F S , Audah L , Hassan M H , et al. A Concise Review of Deep Learning Deployment in 3D Computer Vision Systems[C]// 2021 4th International Symposium on Agents, Multi-Agent Systems and Robotics (ISAMSR). IEEE, 2021:157-160.

[10] Pandey S K , Shekhawat H S , Prasanna S . Deep Learning Techniques for Speech Emotion Recognition : A Review[C]// 29th International Conference Radioelektronika (RADIOELEKTRONIKA). IEEE, 2019:1-6.

[11] Deng L, Yu D. Deep Learning: Methods and Applications[J]. Foundations & Trends in Signal Processing, 2014, 7(3):197-387.

[12] Guan G , Huang H , Song Y , et al. Deep Learning for an Effective Nonorthogonal Multiple Access Scheme[J]. IEEE Transactions on Vehicular Technology, 2018, 67:8440-8450.

[13] Gruber T, Cammerer S, Hoydis J, et al. On Deep Learning-Based Channel Decoding[C]//2017 51st Annual Conference on Information Sciences and Systems (CISS). IEEE, 2017:1-6.

[14] Nachmani E, Be'ery Y, Burshtein D. Learning to Decode Linear Codes Using Deep Learning[C]//2016 54th Annual Allerton Conference on Communication, Control, and Computing (Allerton). IEEE, 2016:341-346.

[15] Goldie C M, Pinch R G. Communication Theory[M]. Cambridge University Press, 1991.

[16] Kaleh G K. Channel Equalization for Block Transmission Systems[J]. IEEE Journal on Selected Areas in Communications, 1995, 13(1):110-121.

[17] 陈良明, 韩泽耀. OFDM-第四代移动通信的主流技术[J].计算机技术与发展,2008,(03):184-187.

[18] 于群,曹娜. MATLAB/Simulink电力系统建模与仿真[M]. 机械工业出版社，2013.

[19] Wu S , Bar-Ness Y . OFDM Channel Estimation in the Presence of Frequency Offset and Phase Noise[C]// IEEE International Conference on Communications. IEEE, 2003:3366-3370.

[20] Raghavendra M R, Giridhar K. Improving Channel Estimation in OFDM Systems for Sparse Multipath Channels[J]. IEEE Signal Processing Letters, 2004, 12(1): 52-55.

[21] 孙晨. 适用于IEEE802.11ac标准的信道估计与均衡技术研究[D]. 重庆: 重庆大学, 2017.

[22] 叶家恒. MIMO-OFDM系统信道估计与检测算法的研究[D]. 广州: 华南理工大学, 2013.

[23] Park H R. Adaptive Channel Estimation for OFDM Systems Using CIR Length Estimate[J]. IEEE Wireless Communications Letters, 2021, 10(11):2597-2601.

[24] Wu H. LMMSE Channel Estimation in OFDM Systems: A Vector Quantization Approach[J]. IEEE Communications Letters, 2021, 25(6):1994-1998.

[25] 郭庆, 李昂迪, 吴景等. 基于ZoomFFT的水声OFDM信号解调算法[J].系统仿真学报,2022,34(02):314-322.

[26] Xiang L, Maunder R G, Hanzo L . Concurrent OFDM Demodulation and Turbo Decoding for Ultra Reliable Low Latency Communication[J]. IEEE Transactions on Vehicular Technology, 2020, 69(2):1281-1290.

[27] Yoon E, Kwon S, Yun U, et al. LDPC Decoding With Low Complexity for OFDM Index Modulation[J]. IEEE Access, 2021, 9:68435-68444.

[28] Deng X, Sha J, Zhou X, et al. Joint Detection and Decoding of Polar-Coded OFDM-IDMA Systems[J]. IEEE Transactions on Circuits and Systems I: Regular Papers, 2019, 66(10):4005-4017.

[29] 万菁晶, 陆怡琪, 田梦倩等. 面向5G无线通信系统中若干物理层技术探讨[J]. 太赫兹科学与电子信息学报, 2018,16(06):962-969.

[30] Huang H , Guo S , Gui G, et al. Deep Learning for Physical-Layer 5G Wireless Techniques: Opportunities, Challenges and Solutions[J].IEEE Transactions on Wireless Communication, 2020, 27(5):126–132.

[31] Peng Q, Li J, Shi H. Deep Learning Based Channel Estimation for OFDM Systems With Doubly Selective Channel[J]. IEEE Communications Letters, 2022, 26(9):2067-2071.

[32] Sun Y, Shen H, Du Z, et al. ICINet: ICI-Aware Neural Network Based Channel Estimation for Rapidly Time-Varying OFDM Systems[J]. IEEE Communications Letters, 2021, 25(9):2973-2977.

[33] Ge L, Guo Y, Zhang Y, et al. Deep Neural Network Based Channel Estimation for Massive MIMO-OFDM Systems with Imperfect Channel State Information[J]. IEEE Systems Journal, 2021, 16(3):4675-4685.

[34] Li J, Peng Q. Lightweight Channel Estimation Networks for OFDM Systems[J]. IEEE Wireless Communications Letters, 2022, 11(10):2066-2070.

[35] Liu S, Huang X. Sparsity-aware channel estimation for mmWave massive MIMO: A deep CNN-based approach [J]. China Communications, 2021, 18(6):162-171.

[36] Ye H , Li G Y , Juang B . Power of Deep Learning for Channel Estimation and Signal Detection in OFDM Systems.[J]. IEEE Wireless Communication. Letters, 2018, 7(1) :114-117.

[37] Wang S, Yao R, Tsiftsis T A, et al. Signal Detection in Uplink Time-Varying OFDM Systems Using RNN with Bidirectional LSTM[J]. IEEE Wireless Communications Letters, 2020, 9(11):1947-1951.

[38] Zheng S, Chen S, Yang X. DeepReceiver: A Deep Learning-Based Intelligent Receiver for Wireless Communications in the Physical Layer[J]. IEEE Transactions on Cognitive Communications and Networking, 2020, 7(1):5-20.

[39] Zhu Y, Wang B, Li J, et al. Y-Shaped Net-Based Signal Detection for OFDM-IM Systems [J]. IEEE Communications Letters, 2022, 26(11):2661-2664.

[40] Ji X , Wang J , Li Y , et al. Modulation Recognition in Maritime Multipath Channels: A Blind Equalization-aided Deep Learning Approach[J]. China Communications, 2020, 17(3):12-25.

[41] Liu Y, Liu Y, Yang C. Modulation Recognition with Graph Convolutional Network[J]. IEEE Wireless Communications Letters, 2020, 9(5):624-627.

[42] Tunze G B, Huynh-The T, Lee J M, et al. Sparsely Connected CNN for Efficient Automatic Modulation Recognition[J]. IEEE Transactions on Vehicular Technology, 2020, 69(12):15557-15568.

[43] An Z, Zhang T, Ma B, et al. Blind High-Order Modulation Recognition for Beyond 5G OSTBC-OFDM Systems via Projected Constellation Vector Learning Network[J]. IEEE Communications Letters, 2021, 26(1):84-88.

[44] Irawan A, Witjaksono G, Wibowo W K. Deep Learning for Polar Codes over Flat Fading Channels[C]//2019 International Conference on Artificial Intelligence in Information and Communication (ICAIIC). IEEE, 2019:488-491.

[45] Guo X, Chang T H, Wang Y. Model-Driven Deep Learning ADMM Decoder for Irregular Binary LDPC Codes[J]. IEEE Communications Letters, 2023, 27(2):571-575.

[46] Wang Q, Liu Q, Wang S, et al. Normalized Min-Sum Neural Network for LDPC Decoding[J]. IEEE Transactions on Cognitive Communications and Networking, 2023, 9(1):70-81.

[47] Zhang Z, Wu F, Lee W S. Factor Graph Neural Networks[J]. Advances in Neural Information Processing Systems, 2020, 33:8577-8587.

[48] Satorras V G, Welling M. Neural Enhanced Belief Propagation on Factor Graphs[C]//International Conference on Artificial Intelligence and Statistics. PMLR, 2021:685-693.

[49] O'Shea T, Hoydis J. An Introduction to Deep Learning for the Physical Layer[J]. IEEE Transactions on Cognitive Communications and Networking, 2017, 3(4) :563-575.

[50] 牟迪，蒙文，赵尚弘等. 基于Wasserstein生成对抗网络的智能光通信[J]. 中国激光, 2020,47(11):224-230.

[51] An Y, Wang S, Zhao L, et al. A Learning-Based End-to-End Wireless Communication System Utilizing a Deep Neural Network Channel Module[J]. IEEE Access, 2023, 11:17441-17453.

[52] 孙志国，申丽然，郭佩等编著. 无线通信链路中的现代通信技术[M]. 北京：电子工业出版社，2010.

[53] Bhavana B, Namburu S, Panigrahi T, et al. Robust Methods for Wideband Compressive Spectrum Sensing Under Non-Gaussian Noise[J]. IEEE Communications Letters, 2021, 25(10):3398-3402.

[54] Freitas M , Egan M , Clavier L , et al. Capacity Bounds for Additive Symmetric α-Stable Noise Channels[J]. IEEE Transactions on Information Theory, 2017, PP(8):1-1.

[55] 张驰, 郭媛, 黎明. 人工神经网络模型发展及应用综述[J].计算机工程与应用, 2021,57(11):57-69.

[56] Huang G, Liu Z, Laurens V D M, et al. Densely Connected Convolutional Networks[J]. Computer Vision and Pattern Recognition, 2017, 1:2261-2269.

[57] He K, Zhang X, Ren S, et al. Deep Residual Learning for Image Recognition[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2016:770-778.

[58] Howard A, Sandler M, Chu G, et al. Searching for Mobilenetv3[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. 2019:1314-1324.

[59] Howard A G, Zhu M, Chen B, et al. Mobilenets: Efficient Convolutional Neural Networks for Mobile Vision Applications[J].2017.

[60] Sandler M, Howard A, Zhu M, et al. MobileNetV2: Inverted Residuals and Linear Bottlenecks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018:4510-4520.

[61] COST207. Digital Land Mobile Radio Communications[J]. Final report of the COST-Project 207, 1998.

[62] Han K , Wang Y , Tian Q , et al. GhostNet: More Features From Cheap Operations[C]// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2020:1577-1586.

[63] Wang Q, Wu B, Zhu P, et al. ECA-Net: Efficient Channel Attention for Deep Convolutional Neural Networks[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2020: 11534-11542.

[64] 高淑萍, 赵清源, 齐小刚等. 改进MobileNet的图像分类方法研究[J]. 智能系统学报, 2021,16(01):11-20.

[65] Misra D . Mish: A Self Regularized Non-Monotonic Activation Function.[C]// British Machine Vision Conference. 2020.

[66] Kysti P , Meinil J , Hentil L , et al. WINNER II channel models[J]. Ero Dk, 2008.