云南大学信息学院;
毫米波通信中大规模多输入多输出(MIMO)系统的实现需要使用大量射频链，导致了系统硬件成本和能耗过高的问题。为了解决这一问题，考虑利用透镜天线阵列基于方向的能量聚焦特性并结合天线选择技术能够在性能损失不明显的同时有效地减少射频链的使用数量。针对具有透镜天线阵列的多用户MIMO系统，提出了一种基于卷积神经网络(CNN)的天线选择算法，将信道状态信息和以信道容量为指标而制作的标签作为网络的输入，对神经网络进行训练，用训练好的网络模型为新的信道状态信息选择出最优的天线组合。仿真结果表明，所提方案获得的和速率性能接近全数字方案，并且所提天线选择算法的分类准确率可以达到98%左右，优于其他基于机器学习的天线选择算法。
| 240 | 1 | 15 |
| 下载次数 | 被引频次 | 阅读次数 |
[1] HAN S F, CHIH-LIN I, XU Z K, et al. Large-scale Antenna Systems with Hybrid Analog and Digital Beamformingfor MillimeterWave 5G[J]. IEEE Communications Magazine, 2015, 53(1):186-194.
[2] GAO X Y, DAI L L, HAN S F, et al. Energy-efficient HybridA nalog and Digital Precoding for mmWave MIMO Systems with Large Antenna Arrays[J]. IEEE Journal on SelectedAreas in Communications, 2016, 34(4):998-1009.
[3] HUANG W, HUANG Y M, ZENG Y, et al. Wideband Millimeter Wave Communication with LensAntenna Array:Joint Beamforming and Antenna Selection with Group SparseOptimization[J]. IEEE Transactions on Wireless Comm unications, 2018, 17(10):6575-6589.
[4] ZENGY, ZHANG R, CHEN Z N. Electromagnetic Lensfocusing Ant enna Enabled Massive MIMO:Performance Improve mentand Cost Reduction[J]. IEEE Journal on Selected Areas in Communications,2014,32(6):1194-1206.
[5] ZHE NGY, ZHANG R. Millimeter Wave MIMO with LensAntenna Array:A New Path Division Multiplexing Paradi gm[J]. IEEE Transactions on Communications,2016,64(4):1557-1571.
[6] JOUNG J. Machine Learning-based Antenna Selection in Wireless Communications[J]. IEEE Communications Lett ers,2016, 20(11):2241-2244.
[7] HED X, LIU C X, QUEK T, et al. Antenna Selection in MIMO Wiretap TransmChannels:A itMachine Learning Approach[J]. IEEE Wireless Communications Lett ers, 2018,7(4):634-637.
[8] LINB, GAO FF, ZHANG S, et al. Deep Learning Base dAntenna Selection and CSI Extrapolation in MassiveMIMO Systems[J]. IEEE Transactions on Wireless Communications, 2021, 20(11):7669-7681.
[9] DIN G Z G, FAN P Z, POOR H V. Random Beamformingin Millimeter-wave NOMA Networks[J]. IEEE Access,2017, 5:7667-7681.
[10] WANG B C, DAI L L, WANG Z C, et al. Sp ectrum and EnergyEfficient Beamspace MIMO-NOMA forMillimeterwave Communications Using Lens Antenna Array[J].IEEE Journal on Selected Areas in Communications,2017,35(10):2370-2382.
[11] AMADORIP V, MASOUROS C. Low RF-complexity Millimeter-wave Beamspace-MIMO Systems by Beam Selecti on[J]. IEEE Transactions on Communications,2015,6 3(6):2212-2223.
[12] ZHANGY H, FAN X, HUO Y, et al. An Improved Offgr id Channel Estimation Scheme for Lens Antenna Arrays inmm Wave Systems[J]. Procedia Computer Science,2020,174:555-560.
[13] GAO X Y, DAI L L, CHEN Z J, et al. Near-optimal Beam Selection for Beamspace mmWave Massive MIMO Systems[J]. IEEECommunications Letters, 2016, 20(5):1054-1057.
[14] ELBIR A M, MISHRA K V, ELDAR Y C. CNN-based Cognitive Radar Array Selection[C]∥2019 IEEE Radar Conference. Boston:IEEE, 2019:1-5.
[15] RIY AZ B, GANAPATHY S. A Deep Learning Approach forEf fective Intrusion Detection in Wireless Networks UsingCNN[J]. Soft Computing,2020,24(22):17265-17278.
[16] LECU N Y, BENGIO Y, HINTON G. Deep Learning[J].Nature, 2015, 521:436-444.
[17] TINO P. Equilibria of Iterative Softmax and Critical Tempe ratures for Intermittent Search in Self-organizing NeuralNetworks[J]. Neural Computation, 2007, 19(4):1056-1081.
[18] LEW ICKI G, MARINO G. Approximation by Superpositions of a Sigmoidal Function[J]. Zeitschrift Für Analysis UndIhre Anwendungen, 2003(22):463-470.
[19] BOUTH ILLIER X, KONDA K, VINCENT P, et al.Dropoutas Data A ugmentation[J/OL].(2015-06-29)[2022-10-24].https:∥arxiv. org/abs/1506. 08700.
[20] KINGMA D P, BA J. Adam:A Method for Stochastic Optimization[J/OL].(2014-12-22)[2022-10-24].https:∥arxiv. org/abs/1412. 6980v6.
基本信息:
DOI：
中图分类号:TN929.5
引用信息:
[1]程文铭,钱蓉蓉,任文平.一种面向透镜阵列毫米波多用户MIMO的CNN天线选择算法[J].无线电工程,2023,53(05):1145-1152.
基金信息:
国家自然科学基金青年科学基金项目(61701433); 云南省科技厅面上项目(2018FB099)~~