多输入多输出技术（Multiple-Input-Multiple-Output，MIMO）是第五代移动通信系统和无线局域网等通信系统的关键技术之一。但是在一些复杂场景下，如煤矿井下受限空间的无线通信环境复杂，信道模型难以精确建立，严重影响着MIMO无线通信系统的性能。然而机遇与挑战并存，人工智能技术的快速发展为解决复杂场景通信中存在的难题提供了新思路，矿井这类复杂场景下的无线通信问题可能更适合采用智能通信的方法来解决。因此，论文基于深度学习，研究了复杂场景MIMO通信智能接收方法，提高了复杂场景无线通信系统的可靠性。主要工作有以下几个方面：

（1）提出了基于深度学习的MIMO通信智能接收方法框架，该方法通过卷积神经网络对接收机的信号处理模块进行整体优化，替代了传统MIMO通信系统接收机的各个模块优化的过程，避免了各个模块优化造成的误差累积，提升了接收机信息恢复的可靠性。在此基础上，针对矿井MIMO通信智能接收方法的实现架构，研究了卷积神经网络，设计了基于密集连接网络的卷积神经网络结构，实现了MIMO通信智能接收方法。在不同信道条件、不同天线数目以及不同神经网络结构下对该方法进行了性能分析。仿真结果表明相较于传统矿井MIMO通信，该方法具有更优的误比特率性能。

（2）为了便于矿井MIMO通信智能接收方法在终端实现，又进一步研究了基于轻量型卷积神经网络的智能接收方法的实现架构，设计了基于MobileNetV2和MobileNetV3的卷积神经网络模型，引入一个浅层特征提取模块，并借鉴了DenseNet中密集连接的思想。考虑了输入信息比特流位数的不同，改进了原始网络的分类层，实现了便于终端部署的MIMO通信智能接收方法。在不同调制方式、不同输入信息比特流位数以及不同神经网络结构下进行了性能分析。仿真结果表明该方法降低了模型的训练和预测时间，但仍能提高矿井MIMO通信的性能。

       复杂场景MIMO通信智能接收方法发挥了深度学习技术从数据中获得更深层次特征的优势，克服了矿井通信环境对接收机信息恢复的影响，改善了矿井MIMO通信的性能。该方法不仅适用于矿井等地下受限空间的无线通信，也对其他复杂场景的无线通信具有一定的借鉴作用。

      Multiple-Input-Multiple-Output (MIMO) technology is one of the key technologies for the fifth generation mobile communication systems and wireless local area networks. However, in some complex scenes, such as the complex wireless communication environment in the confined space of underground coal mine, the channel model is difficult to establish accurately, which seriously affects the performance of the MIMO wireless communication system. However, opportunities and challenges coexist. The rapid development of artificial intelligence technology provides new ideas for solving problems in complex communication scenes. Wireless communication problems in complex scenes like mine may be more suitable to be solved by intelligent communication. Therefore, based on deep learning, this thesis studies the intelligent receiving method for MIMO communication under complex scenes, which improves the reliability of the wireless communication system. The main work includes the following aspects:
      (1)A deep learning based intelligent receiving method for MIMO communication is proposed. This method optimizes the signal processing module of the receiver as a whole through a convolutional neural network, which replaces each module of the traditional MIMO communication receiver, avoids the accumulation of errors caused by modular operation and improves the reliability of information recovery at the signal receiving end. On this basis, aiming at implementation architecture of the intelligent receiving methods for mine MIMO communication, the convolutional neural network is studied and the convolutional neural network model based on DenseNet is designed. Then the intelligent receiving method for MIMO communication is realized. The performance of the method is analyzed under different channel conditions, different numbers of antennas and different neural network structures. Simulation results show that compared with the traditional mine MIMO communication, the proposed method has better bit error rate performance.
      (2)In order to facilitate the implementation of the intelligent receiving method for mine MIMO communication in the terminal, a lightweight convolutional neural network is further studied, the convolutional neural network model based on MobileNetV2 and MobileNetV3 is designed. A shallow feature extraction module is introduced, and the idea of dense connection in DenseNet is borrowed. Considering the difference of input bit, the classification layer of original network is improved, and the intelligent receiving method of MIMO communication is realized which is convenient for terminal. The performance is analyzed under different modulation methods, different bit stream bits of input information and different neural network structures. Simulation results show that the lightweight convolutional neural networks reduce the training and predict time of the model, while still improve the performance of mine MIMO communication.
       The intelligent receiving method of MIMO communication in complex scenes takes advantage of deep learning technology to obtain deeper features from data, overcomes the influence of mine communication environment on information recovery, and improves the performance of mine MIMO communication. This method is not only suitable for wireless communication in underground confined space such as mine, but also can be used as a reference for wireless communication in other complex scenes.

[1] Yang S, Hanzo L. Fifty Years of MIMO Detection: The Road to Large-Scale MIMOs[J]. IEEE Communications Surveys & Tutorials, 2017, 17(4):1941-1988.

[2] Larsson, Erik, Edfors, et al. Massive MIMO for next generation wireless systems.[J]. IEEE Communications Magazine, 2014, 52(2) : 186-195.

[3] Huang D, Jiang X. -Q, Chen S, et al. Deep Learning Based Parallel Detector for MIMO Systems[C]//. Proceedings of the 5th International Conference on Communication, Image and Signal Processing (CCISP 2020), 2020:46-50.

[4] 王国法, 赵国瑞, 胡亚辉. 5G技术在煤矿智能化中的应用展望[J].煤炭学报, 2020,45(01):16-23.

[5] 孙继平. 煤矿智能化与矿用5G[J].工矿自动化,2020,46(08):1-7.

[6] 姚善化. 基于镜像法的矿井隧道电磁波多径信道模型[J].工矿自动化, 2017,43(04):46-49.

[7] 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, 2018,34(08):46-55.

[8] 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.

[9] 李国权, 徐永海, 林金朝, 徐勇军, 杨鹏. 基于深度学习的无线物理层关键技术研究综述[J].重庆邮电大学学报(自然科学版), 2020,32(04):503-510.

[10]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. vol. 27, no. 5, pp. 126–132, 2020

[11]Gui G, Liu M, Tang F, et al. 6G: Opening New Horizons for Integration of Comfort, Security and Intelligence[J]. IEEE Wireless Communications, 2020, 27(5) : 1-7.

[12]You X, Wang C, Huang J, et al .Towards 6G wireless communication networks:vision, enabling technologies, and new paradigm shifts[J].Science China(Information Sciences),2021,64(01):5-78.

[13]Mabrouk I.B, Talbi L, Nedil M. Enhancement of wireless channel performance using MIMO system in underground mine gallery[C]//. 2012 International Conference on Wireless Communications in Underground and Confined Areas, Clermont Ferrand, 2012, pp. 1-5.

[14]冯振杰. 多天线环境下的无线协作通信技术[M]. 北京：科学技术文献出版社，2016.

[15]Alamouti S.M. A simple transmit diversity technique for wireless communications[J]. IEEE Journal on Selected Areas in Communications, 1998, 16(8) : 1451-1458.

[16]Tarokh V, Calderbank A.R, Jafarkhani H. Space-time block codes from orthogonal designs[J]. IEEE Transactions on Information Theory, 1999, 45(5) : 1456-1467.

[17]孟宪猛, 沈磊, 姚善化. 基于MIMO技术的矿井无线通信系统的设计[J].煤炭工程, 2011(03):122-124.

[18]Mabrouk B, Talbi L, Nedil M. Performance Evaluation of a MIMO System in Underground Mine Gallery[J]. IEEE Antennas and Wireless Propagation Letters, vol. 11, pp. 830-833, 2012.

[19]Mabrouk B, Talbi L, Nedil M, et al. MIMO-UWB Channel Characterization Within an Underground Mine Gallery[J]. IEEE Transactions on Antennas and Propagation, vol. 60, no. 10, pp. 4866-4874, Oct. 2012.

[20]Taleb H.A, Nedil M, Ghanem K, et al. MIMO-OFDM performance evaluation over meausured underground mine channel at 2.4 GHz[C]// 2016 16th Mediterranean Microwave Symposium (MMS), 2016, pp. 1-4.

[21]Taleb H, Ghanem K, Zaaimia M.Z, et al. On 60 GHz MIMO Diversity in an Underground Mine Propagation Channel[J]. IEEE Antennas and Wireless Propagation Letters, vol. 19, no. 10, pp. 1769-1773, Oct. 2020.

[22]Labdaoui R, Ghanem K, Youcef Ettoumi F, et al. Enhancement of MIMO-UWB Communications in Underground Mine Environments[C]// 2021 IEEE International Symposium on Antennas and Propagation and USNC-URSI Radio Science Meeting (APS/URSI), 2021, pp. 539-540.

[23]Bai Q, Wang J, Zhang Y, et al. Deep Learning-Based Channel Estimation Algorithm Over Time Selective Fading Channels[J]. IEEE Transactions on Cognitive Communications and Networking, 2020, 6(1):125-134.

[24]Mashhadi M. B, Gündüz D. Pruning the Pilots: Deep Learning-Based Pilot Design and Channel Estimation for MIMO-OFDM Systems[J]. IEEE Transactions on Wireless Communications, vol. 20, no. 10, pp. 6315-6328, Oct. 2021.

[25]Cheng X, Liu D, Wang C, et al. Deep Learning-Based Channel Estimation and Equalization Scheme for FBMC/OQAM Systems.[J]. IEEE Wireless Commun. Letters, 2019, 8(3) : 881-884.

[26]Huang H, 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, vol. 67, no. 9, pp. 8549-8560, Sept. 2018.

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

[28]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, PP(99):1-1.

[29]Samuel N, Diskin T, Wiesel A. Deep MIMO Detection[C]// 2017 IEEE 18th International Workshop on Signal Processing Advances in Wireless Communications (SPAWC). arXiv, 2017.

[30]Liao J, Zhao J, Gao F et al. A Model-Driven Deep Learning Method for Massive MIMO Detection[J]. IEEE Communications Letters, vol. 24, no. 8, pp. 1724-1728, Aug. 2020.

[31]Samuel N, Diskin T, Wiesel A. Learning to Detect[J]. Signal Processing, IEEE Transactions on, 2019.

[32]吴佩军, 侯进, 吕志良等. 基于卷积神经网络的多进制相位调制信号识别算法[J].计算机应用与软件, 2019,36(11):202-209.

[33]张思成, 林云, 涂涯等. 基于轻量级深度神经网络的电磁信号调制识别技术[J].通信学报, 2020,41(11):12-21.

[34]Xie X, Yang G, Jiang M, et al. A Kind of Wireless Modulation Recognition Method Based on DenseNet and BLSTM[J]. IEEE Access, vol. 9, pp. 125706-125713, 2021.

[35]Shah H, Dang X. Low-complexity deep learning and RBFN architectures for modulation classification of space-time block-code (STBC)-MIMO systems[J]. Digital Signal Processing,Volume 99,2020.

[36]Fei L, Cong S, Feng W. An Iterative BP-CNN Architecture for Channel Decoding[J]. IEEE Journal of Selected Topics in Signal Processing, 2018, 12(1) : 144-159.

[37]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), Okinawa, Japan, 2019, pp. 488-491.

[38]Gruber T, Cammerer S, Hoydis J. On deep learning-based channel decoding.[C]// 2017 51st Annual Conference on Information Sciences and Systems (CISS), 2017, pp. 1-6.

[39]Zhang J, He H, Yang X et al. Model-driven Deep Learning Based Turbo-MIMO Receiver[C]// 2020 IEEE 21st International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), 2020, pp. 1-5.

[40]Dorner S, Cammerer S, Hoydis J. et al. Deep Learning Based Communication Over the Air[J]. IEEE Journal of Selected Topics in Signal Processing, 2018, 12(1) : 132-143.

[41]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.

[42]李国权, 杨鹏, 马珺杰, 徐永海, 林金朝. 基于深度学习的MIMO系统联合信号检测与信道译码[J].重庆邮电大学学报(自然科学版), 2021,33(02):176-184.

[43]哈米德·贾法哈尼. 空时编码的理论与实践[M].西安：西安交通大学出版社，2007.

[44]Murthy N S, Gowri S S, Rao B P. Novel orthogonal space time block codes for multiple transmit antennas(nine and ten antennas with full code rate)[C]// 2015 2nd International Conference on Electronics and Communication Systems (ICECS), Coimbatore, 2015, pp. 675-678.

[45]陈超, 齐峰. 卷积神经网络的发展及其在计算机视觉领域中的应用综述[J].计算机科学, 2019, 46(3): 63–73.

[46]严春满, 王铖. 卷积神经网络模型发展及应用[J].计算机科学与探索, 2021,15(01):27-46.

[47]张玉宏. 深度学习之美[M].北京：电子工业出版社, 2018.

[48]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.

[49]林景栋, 吴欣怡, 柴毅, 尹宏鹏. 卷积神经网络结构优化综述[J].自动化学报, 2020,46(01):24-37.

[50]Li Z, Liu F, Yang W, et al. A Survey of Convolutional Neural Networks: Analysis, Applications, and Prospects.[J]. IEEE transactions on neural networks and learning systems, 2021.

[51]杨维, 李滢, 孙继平. 类矩形矿井巷道中UHF宽带电磁波统计信道建模[J].煤炭学报, 2008(04):467-472.

[52]张玉, 杨维. 矿井巷道综合衰减模型与无线传输误码率性能综合比较[J].煤炭学报, 2011,36(09):1581-1586.

[53]王艳芬, 于洪珍 ,张传祥. 矿井超宽带复合衰落信道建模及仿真[J].电波科学学报, 2010,25(04):805-812.

[54]Ranjan A, Misra P, Sahu H B. Experimental measurements and channel modeling for wireless communication networks in underground mine environments[C]//2017 11th European Conference on Antennas and Propagation (EUCAP). IEEE, 2017: 1345-1349.

[55]Wang B, Xu K, Zheng S, et al. A Deep Learning-Based Intelligent Receiver for Improving the Reliability of the MIMO Wireless Communication System.[J] IEEE Transactions on Reliability, vol. 71, no. 2, pp. 1104-1115, June 2022.

[56]Huang G, Liu Z, Laurens V, et al. Densely Connected Convolutional Networks[J]. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, 2017, pp. 2261-2269.

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

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

[59]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.

[60]Howard A, Sandler M, Chen B, et al. Searching for MobileNetV3[C]// 2019 IEEE/CVF International Conference on Computer Vision (ICCV), 2019, pp. 1314-1324.

[61]Hu J, Shen L, Albanie S, et al. Squeeze-and-Excitation Networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(8):2011–2023.

[62]Kingma D, Ba J. Adam: A Method for Stochastic Optimization[J]. Computer Science, 2014.