Aiming at the problems of traditional information symbol recognition methods relying on a series of signal processing at the receiver backend and the difficulties in obtaining training data in real time and the high cost of computing resources in deep learning based the information symbol recognition methods, a single-stream information symbol recognition method based on a programmable diffraction deep neural network (PD2NN) is proposed. Firstly, the problem of information symbol recognition is transformed into the recognition of electromagnetic waves at the receiving antenna, and a single-stream information symbol recognition model based on PD2NN is proposed. Secondly, a PD2NN parameter training method based on adaptive moment estimation (ADAM) is designed. Finally, the PD2NN network is used to recognize the single stream information symbol by judging the type of information symbol with the smallest distance from the received symbol. Simulation results show that the bit error rate of the proposed method is comparable to or even better than that of the traditional receiver method, and the symbol recognition of single-stream information under the condition of lower training cost can be achieved.
由大量可配置无源电磁元件组成的信息超材料[11-12],能够在电磁波的传播过程中调控其振幅、相位和极化等特性。作为信息超材料的一种,数字编码超表面与深度神经网络中的网络层结构相似。基于此,现有研究利用多层透射数字编码超表面实现可编程衍射神经网络(Programmable Diffractive Deep Neural Network, PD2NN)架构[13]。文献[13]中将携带不同种类图像信息的电磁波经过PD2NN处理后,在输出层通过检测对应端口的电磁波能量分布识别出图像,表明PD2NN在电磁域执行电磁波计算的优势和实现模式识别的能力。同样从场波域的视角审视无线通信中的信号传播,携带不同信息符号的电磁波具有不同的电磁特征。由此,探索利用PD2NN执行深度学习的优势进行信息符号识别,有望突破现有信息符号识别方法面临的瓶颈。
基于上述思想,本文面向单流信号通信场景,开展PD2NN在数字调制信息符号识别领域的探索性研究,设计一种基于PD2NN的单流信息符号识别方法。首先分析了单流信息符号在电磁域的可分性,将对信息符号的识别问题转换为对接收天线处电磁波的识别问题。其次,利用PD2NN分类不同特征电磁波的能力,提出基于PD2NN的单流信息符号识别模型。再次,提出基于自适应矩估计(Adaptive Moment Estimation, ADAM)的PD2NN参数训练方法,通过训练使其具备识别信息符号的能力。最后,PD2NN网络通过判断与接收符号距离最小的信息符号类别,实现对单流信息符号的识别。仿真结果表明,本文所提方法具有较好的误码率性能。
PROAKISJ G, SALEHIM. Digital communications [M]. New York, USA: McGraw-Hill, 2008:160-226.
[2]
LIH C. Application analysis of digital signal processing technology in 5G communication[C]∥Proceedings of the 2023 Asia-Europe Conference on Electronics, Data Processing and Informatics. Piscataway, USA: IEEE, 2023:476-481.
[3]
WANGJ, LIR, WANGJ, et al. Artificial intelligence and wireless communications[J]. Frontiers of Information Technology & Electronic Engineering, 2020,21(10):1413-1425.
[4]
MAHMOUDH, ELBADAWYH M, ISMAILT, et al. A comprehensive review of generative AI applications in 6G[C]∥Proceedings of the 2024 6th Novel Intelligent and Leading Emerging Sciences Conference. Piscataway, USA: IEEE, 2024:593-596.
[5]
SAIMLERM, İCKINS, BERNINIG, et al. The role of AI enablers in overcoming impairments in 6G networks[C]∥Proceedings of the 2024 Joint European Conference on Networks and Communications & 6G Summit. Piscataway, USA: IEEE, 2024:913-918.
[6]
O’SHEAT, HOYDISJ. An introduction to deep learning for the physical layer[J]. IEEE Transactions on Cognitive Communications and Networking, 2017,3(4):563-575.
[7]
DÖRNERS, CAMMERERS, HOYDISJ, et al. Deep learning based communication over the air[J]. IEEE Journal of Selected Topics in Signal Processing, 2018,12(1):132-143.
[8]
TONGW, LIG Y. Nine challenges in artificial intelligence and wireless communications for 6G[J]. IEEE Wireless Communications, 2022,29(4):140-145.
[9]
WUN, JIANGR K, WANGX Y, et al. AI-enhanced integrated sensing and communications: advancements, challenges, and prospects[J]. IEEE Communications Magazine, 2024,62(9):144-150.
[10]
ELSAYEDM, EROL-KANTARCIM. AI-enabled future wireless networks: challenges, opportunities, and open issues[J]. IEEE Vehicular Technology Magazine, 2019,14(3):70-77.
[11]
MAQ, CUIT J. Information metamaterials: bridging the physical world and digital world[J]. PhotoniX, 2020,1:No.1.
[12]
LIL L, CUIT J. Information metamaterials: from effective media to real-time information processing systems[J]. Nanophotonics, 2019,8(5):703-724.
[13]
LIUC, MAQ, LUOZ J, et al. A programmable diffractive deep neural network based on a digital-coding metasurface array[J]. Nature Electronics, 2022,5(2):113-122.
[14]
ISHIMARUA. Electromagnetic wave propagation, radiation, and scattering: from fundamentals to applications[M]. Hoboken, USA: Wiley-IEEE Press, 2017:137-169.
[15]
沙威,应磊,肖高标.电磁学中的格林函数[J].物理,2024,53(1):43-49.
[16]
LINX, RIVENSONY, YARDIMCIN T, et al. All-optical machine learning using diffractive deep neural networks[J]. Science, 2018,361(6406):1004-1008.
[17]
ANJ C, YUENC, GUANY L, et al. Stacked intelligent metasurface performs a 2D DFT in the wave domain for DOA estimation[C]∥Proceedings of the 2024 IEEE International Conference on Communications. Piscataway, USA: IEEE, 2024:3445-3451.
[18]
ZHANGZ J. Improved Adam optimizer for deep neural networks[C]∥Proceedings of the 2018 IEEE/ACM 26th International Symposium on Quality of Service. Piscataway, USA: IEEE, 2018. DOI: 10.1109/IWQoS.2018.8624183 .
[19]
LIUW B, WANGZ D, LIUX H, et al. A survey of deep neural network architectures and their applications[J]. Neurocomputing, 2017,234:11-26.