相较于陆地通信系统，海洋环境复杂且通信设备安装困难制约了海洋通信的发展，我国海洋强国战略的提出对海洋通信系统提出更高的要求。深度学习在无线通信领域中的应用越来越广泛，在此基础上研究基于密集连接网络（Densely Connected Convolutional Networks，DenseNet）的海洋OFDM和FBMC信号检测算法，具体研究内容如下：

（1）针对传统OFDM信号检测算法在复杂海洋信道下误比特率高、对循环前缀和导频的依赖性强的问题，提出基于DenseNet的OFDM信号检测算法，这是一种由网络模型实现的端到端的信号检测算法。根据OFDM信号结构设计DenseNet模型，使用网络模型代替通信系统接收端的全部功能模块。将接收端接收到的IQ信号作为数据集，原始的基带信号作为数据标签训练网络，使用测试集对模型进行仿真验证。在海洋信道下，所提算法具有更好的信号检测性能，且对导频和循环前缀的依赖性较小。

（2）针对海洋环境下传统FBMC信号检测算法计算复杂度高、误比特率高的问题，提出基于改进DenseNet的FBMC信号检测算法。FBMC在每个子载波上使用原型滤波器对信号进行滤波，增加了信号结构的复杂性，也增加了网络模型提取数据集中有用特征的难度，使用SE-Net对原始DenseNet模型进行优化，通过权重分配，增强对训练结果有利的特征，提高模型性能。在高斯白噪声和海洋信道模型下，相比于传统FBMC信号检测算法，基于网络模型的检测算法信号恢复性能更好，误比特率更低，改进后的网络模型可以进一步提高信号检测性能。针对FBMC使用OQAM调制在与MIMO技术结合时会因为系统内部干扰严重影响接收端信号检测性能的问题，使用上述两种网络模型实现FBMC-MIMO接收端的信号检测。实验结果表明，系统内部干扰对网络模型的影响较小，基于网络模型的信号检测算法在FBMC-MIMO系统中具有较好的信号检测性能。

Compared with terrestrial communication systems, the complex marine environment and the difficulty in installing communication equipment restrict the development of marine communication systems. The proposal of our country's maritime power strategy puts forward higher requirements for the marine communication system The application of deep learning in the field of wireless communication is becoming more and more extensive. On this basis, the OFDM and FBMC signal detection algorithms in the marine environment based on densely connected networks are studied. The specific research contents are as follows:

(1) Aiming at the problems of low bit error rate and strong dependence on cyclic prefix and pilot of traditional OFDM in the complex marine channel, an OFDM signal detection algorithm based on DenseNet is proposed, which is an end-to-end signal detection algorithm implemented by a network model. The DenseNet model is designed according to the structure of the OFDM, and it is used to replace all the functional modules of the receiving end of the communication system. The IQ signal received by the receiver is used as the data set, and the original baseband signal is used as the data label to train the network. Then use the test set to simulate and verify the model. Under the marine channel model, the proposed algorithm has better signal detection and recovery performance and less dependence on pilot and cyclic prefix.

(2) Aiming at the problems of high computational complexity and bit error rate of traditional FBMC signal detection algorithm in the complex marine environment, an FBMC signal detection algorithm based on improved DenseNet is proposed. FBMC uses prototype filter on each sub-carrier to filter the signal, which increases the complexity of the signal structure and also increases the difficulty of the network model to extract useful features in the data set. The original DenseNet model is optimized using SE-Net, and the features that are beneficial to the training results are enhanced through weight distribution to improve the model performance. Under the Gaussian white noise and the marine channel models, compared with the traditional FBMC signal detection algorithm, the detection algorithm based on the DenseNet has better signal recovery performance and lower bit error rate. The optimized model can further improve the signal detection performance. Aiming at the problem that the use of OQAM modulation in FBMC combined with MIMO technology will seriously affect the signal detection performance of the receiving end due to the internal interference of the system, the above two network models are used to realize the signal detection of the FBMC-MIMO. The experimental results show that the internal interference of the system has little influence on the network model, and the network model has better signal detection performance in the FBMC-MIMO system.

[1]夏明华, 朱又敏, 陈二虎, 等. 海洋通信的发展现状与时代挑战[J]. 中国科学: 信息科学, 2017, 47(06): 677-695.

[2]陈为刚, 杨晋生, 马建国,等. 海洋通信的最新进展综述[J]. 中国通信, 2012, 9(02): 31-42.

[3]张海君, 苏仁伟, 唐斌, 等. 未来海洋通信网络架构与关键技术[J]. 无线电通信技术, 2021, 47(04): 384-391.

[4]Gao Z, Ke S, Chen C, et al. Ray tracing based marine mobile wireless channel modeling[C]// 2018 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW). IEEE, 2018: 1-2.

[5]史经展, 宗鹏. 关于海上通信多径衰落的仿真研究[J]. 计算机仿真, 2015, 32(12): 152-157+163.

[6]于永学, 王玉珏, 解嘉宇. 海洋通信的发展现状及应用构想[J]. 海洋信息, 2020, 35(02): 25-28.

[7]Banelli P, Colavolpe G, Rugini L, et al. Spectral efficiency of multicarrier schemes for 5G[C]// 2019 26th International Conference on Telecommunications (ICT). IEEE, 2019: 124-129.

[8]魏明君. 面向5G的通用滤波多载波传输方案研究[D]. 南京: 东南大学, 2017.

[9]马宏彪, 徐岩. 高速场景下基于OFDM的信道估计算法[J]. 现代电子技术, 2021, 44(05): 5-10.

[10]李素月, 郑宝玉. 双选择性信道下发射分集MISO OFDM系统基于BEM的低复杂度信道均衡[J]. 信号处理, 2014, 30(11): 1363-1369.

[11]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.

[12]Xu L, Yang J, Huang D, et al. Exploiting cyclic prefix for turbo-OFDM receiver design[J]. IEEE Access, 2017,5: 15762-15775.

[13]Al-Naffouri T Y, Quadeer A A. Cyclic prefix based enhanced data recovery in OFDM[J]. IEEE Transactions on Signal Processing, 2010, 58(6): 3406-3410.

[14]Rathinakumar S, Radunovic B, Marina M K. CPRecycle: Recycling cyclic prefix for versatile interference mitigation in OFDM based wireless systems[C]// Proceedings of the 12th International on Conference on emerging Networking Experiments and Technologies. 2016: 67-81.

[15]熊刚, 彭勃, 曹建辉. 一种改进的OFDM信号检测方法研究[J]. 通信技术, 2018, 51(10): 2309-2313.

[16]Slipenchuk K S, Filimonov A A. The processing of OFDM system signals in dispersion channel[C]// 2020 IEEE Conference of Russian Young Researchers in Electrical and Electronic Engineering (EIConRus). IEEE, 2020: 1428-1431.

[17]李赛峰, 付加飞, 戚婷, 等. OFDM系统群稀疏信道估计与译码迭代算法[J]. 数据采集与处理, 2018,33(06): 986-994.

[18]于柯远, 张立民, 闫文君, 等. 基于特征序列的时域STBC-OFDM盲识别算法[J]. 北京航空航天大学学报, 2021,47(08): 1524-1532.

[19]Ou J, Wang J, Peng Q, et al. Learning-based signal detection for OFDM systems with I/Q imbalance[C]// 2020 IEEE 3rd International Conference on Information Communication and Signal Processing (ICICSP). IEEE, 2020: 69-73.

[20]张岑, 王彪, 朱雨男. 基于深度学习的索引多载波信号检测[J]. 现代电子技术, 2021, 44(05): 80-83.

[21]Kim M, Lee W, Cho D H. A novel PAPR reduction scheme for OFDM system based on deep learning[J]. IEEE Communications Letters, 2018, 22(3):510-513.

[22]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.

[23]常代娜, 周杰. 基于深度学习算法的OFDM信号检测[J]. 东南大学学报(自然科学版), 2020, 50(05): 912-917.

[24]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.

[25]Liang F, Shen C, Wu F. An iterative BP-CNN architecture for channel decoding[J]. IEEE Journal of Selected Topics in Signal Processing, 2018, 12(01): 144-159.

[26]黄媛媛, 张剑, 周兴建, 等 应用深度学习的信号解调[J]. 电讯技术, 2017, 57(07): 741-744.

[27]Yang Y, Gao F, Ma X, et al. Deep learning-based channel estimation for doubly selective fading channels[J], IEEE Access, 2019, 7: 36579-36589.

[28]Raj V, Kalyani S. Backpropagating through the air: deep learning at physical layer without channel models[J]. IEEE Communications Letters, 2018, 22(11): 2278-2281.

[29]Kumari S, Srinivas K K, Kumar P. Channel and carrier frequency offset equalization for OFDM based UAV communications using deep learning[J]. IEEE Communications Letters, 2020, 25(3): 850-853.

[30]Zhao Z, Vuran M C, Guo F, et al. Deep-waveform: A learned OFDM receiver based on deep complex-valued convolutional networks[J]. IEEE Journal on Selected Areas in Communications, 2021, 39(8): 2407-2420.

[31]刘春辉, 王美琳, 董赞亮, 等. 基于调制卷积神经网络的空地数据链信道估计[J]. 北京航空航天大学学报, 2022, 48(03): 533-543.

[32]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(03): 348-353.

[33]王星, 马天鸣, 赵清华. FBMC-OQAM系统中一种改进的OPTS算法[J]. 山东大学学报(工学版), 2021, 51(04): 35-42.

[34]张天骐, 王胜, 李群, 等. 基于相关性的FBMC-OQAM信号的符号周期盲估计[J]. 系统工程与电子技术, 2019, 41(06): 1402-1407.

[35]Li L, Ding L, Wang Y, et al. Joint symbol timing and channel estimation for FBMC-OQAM systems[J]. IEEE Access, 2019, 7: 131326-131337.

[36]徐兰天. 5G系统中低延时的FBMC调制方法[J]. 电子测量技术, 2018, 41(05): 123-127.

[37]Isnawati A F, Citra V O, Hendry J. Performance analysis of audio data transmission on FBMC-offset QAM system[C]// 2019 IEEE international conference on industry 4.0, artificial intelligence, and communications technology (IAICT). IEEE, 2019: 81-86.

[38]Kofidis E. On Preamble-based FBMC/OQAM highly frequency selective channel estimation without guard symbols[C]// 2021 IEEE Global Communications Conference (GLOBECOM). IEEE, 2021: 1-6.

[39]耿云龙, 张瑞. 基于信干噪比的多输入多输出滤波器组多载波系统信道均衡算法[J]. 科学技术与工程, 2019,19(09): 133-137.

[40]Kofidis E. A tensor-based approach to joint channel estimation/data detection in flexible multicarrier MIMO systems[J]. IEEE Transactions on Signal Processing, 2020, 68: 3179-3193.

[41]Singh P, Mishra H B, Jagannatham A K, et al. Semi-blind, training, and data-aided channel estimation schemes for MIMO-FBMC-OQAM systems[J]. IEEE Transactions on Signal Processing, 2019, 67(18): 4668-4682.

[42]徐伟, 张天骐, 冯嘉欣, 等. FBMC-OQAM信号子载波盲估计[J]. 信号处理, 2020, 36(05): 748-755.

[43]Cheng X, Liu D, Wang C, et al. Deep learning based channel estimation and equalization scheme for FBMC/OQAM systems[J]. IEEE Wireless Communications Letters, 2019, 8(03): 881-884.

[44]Chu J, Gao M, Liu X, et al. Channel estimation based on complex-valued neural networks in IM/DD FBMC/OQAM transmission system[J]. Journal of Lightwave Technology, 2022, 40(4): 1055-1063.

[45]何超逸, 袁伟娜. 基于实值神经网络的FBMC/OQAM系统PAPR降低方法[J]. 华东理工大学学报(自然科学版), 2021, 48: 1-7.

[46]朱雨男, 解方彤, 张明亮, 等. 基于多层双向长短时记忆网络的水声索引调制滤波器组多载波系统索引检测方法[J]. 电子与信息学报, 2022, 44: 1-7.

[47]Wei T, Feng W, Ge N, et al. Environment-aware coverage optimization for space-ground integrated maritime communications[J]. IEEE Access, 2020, 8: 89205-89214.

[48]洪东. 基于大数据分析的海上通信网络调度优化算法[J]. 舰船科学技术, 2020, 42(24): 142-144.

[49]陈为刚, 刘元松, 潘州, 等. 面向海上通信的可伸缩广义通信平台架构[J]. 陕西师范大学学报(自然科学版), 2021, 49(01): 88-95.

[50]Hoeft M, Gierlowski K, Rak J, et al. Non-satellite broadband maritime communications for e-navigation services[J]. IEEE Access, 2021, 9: 62697-62718.

[51]Mehrnia N, Ozdemir M K. Novel maritime channel models for millimeter radiowaves[C]// 2016 24th International Conference on Software, Telecommunications and Computer Networks (SoftCOM). IEEE, 2016: 1-6.

[52]赵雨薇, 迟迅, 任佳. 基于ITM模型的海上移动信道传输模型[J]. 电子技术应用, 2014, 40(07): 106-108+111.

[53]魏特, 王文浩, 陈军, 等. 环境信息辅助的海上无线信道测量与建模[J]. 清华大学学报(自然科学版), 2021, 61(09): 1002-1007.

[54]Yang K, Molisch A F, Ekman T, et al. A deterministic round earth loss model for open-sea radio propagation[C]// 2013 IEEE 77th Vehicular Technology Conference (VTC Spring). IEEE, 2013: 1-5.

[55]Wang W, Raulefs R, Jost T. Fading characteristics of maritime propagation channel for beyond geometrical horizon communications in C-band[J]. CEAS Space Journal, 2019, 11(1): 95-104.

[56]Jiang K, Pan D, Jiang T, et al. Ocean surface stochastic channel modeling based on hidden Markov model[C]// 2018 IEEE Asia-Pacific Conference on Antennas and Propagation (APCAP), 2018: 440-441.

[57]Lee J H, Choi J, Lee W H, et al. Measurement and analysis on land-to-ship offshore wireless channel in 2.4 GHz[J]. IEEE Wireless Communications Letters, 2017, 6(2): 222-225.

[58]Yang K, Molisch A F, Ekman T, et al. A round earth loss model and small-scale channel properties for open-sea radio propagation[J]. IEEE Transactions on Vehicular Technology, 2019, 68(9): 8449-8460.

[59]刘西川, 贺彬晟, 印敏, 等. 5G通信中的视距信号大气衰落特性研究[J]. 移动通信, 2018,42(09): 9-15.

[60]Jie H, Li S, Gang S, et al. Squeeze-and-excitation networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020 42(8): 2011-2023.

[61]杨雅涵. 面向5G的广义频分复用调制技术研究[D]. 南京: 东南大学, 2018.