基于深度学习的传感云sink节点最优能效SWIPT波束成形设计

doi:10.11959/j.issn.1000-436x.2021131

通信学报 ›› 2021, Vol. 42 ›› Issue (7): 176-188.doi: 10.11959/j.issn.1000-436x.2021131

基于深度学习的传感云sink节点最优能效SWIPT波束成形设计

王哲¹^,²^,³, 李陶深⁴, 葛丽娜¹^,³, 张桂芬¹, 吴敏⁵^,⁵

¹ 广西民族大学人工智能学院，广西南宁 530006
² 广西混杂计算与集成电路设计分析重点实验室，广西南宁530006
³ 广西民族大学网络通信工程重点实验室，广西南宁 530006
⁴ 广西大学计算机与电子信息学院，广西南宁 530004
⁵ 广西电网有限责任公司科技信息部，广西南宁 530023

修回日期:2021-01-05 出版日期:2021-07-25 发布日期:2021-07-01
作者简介:王哲（1991− ），男，河南南阳人，博士，广西民族大学讲师、硕士生导师，主要研究方向为无线携能通信、传感云系统、机器学习等
李陶深（1957− ），男，广西南宁人，博士，广西大学教授、博士生导师，主要研究方向为移动无线网络、无线能量传输、物联网与智慧城市等
葛丽娜（1969− ），女，广西环江人，博士，广西民族大学教授、硕士生导师，主要研究方向为网络与信息安全、移动计算、人工智能等
张桂芬（1974− ），女，广西南宁人，广西民族大学副教授、硕士生导师，主要研究方向为移动边缘计算、人工智能等
吴敏（1979− ），男，广西南宁人，广西大学博士生，广西电网有限责任公司高级工程师，主要研究方向为电力系统规划、新能源技术、电力电子技术等
基金资助:
国家自然科学基金资助项目(61862007);国家自然科学基金资助项目(62066005);广西自然科学基金资助项目(2020GXNSFBA297103);广西高校中青年教师科研基础能力提升项目(2020KY04030);广西民族大学引进人才科研启动项目(2019KJQD17)

Optimal energy-efficiency beamforming design for SWIPT-enabled sink in sensor cloud based on deep learning

Zhe WANG¹^,²^,³, Taoshen LI⁴, Lina GE¹^,³, Guifen ZHANG¹, Min WU⁵^,⁵

¹ Institute of Artificial Intelligence, Guangxi University for Nationalities, Nanning 530006, China
² Guangxi Key Laboratory of Hybrid Computation and IC Design Analysis, Nanning 530006, China
³ Key Laboratory of Network Communication Engineering, Guangxi University for Nationalities, Nanning 530006, China
⁴ School of Computer, Electronics and Information, Guangxi University, Nanning 530004, China
⁵ Science and Technology Information Department, Guangxi Power Grid Co., Ltd., Nanning 530023, China

Revised:2021-01-05 Online:2021-07-25 Published:2021-07-01
Supported by:
The National Natural Science Foundation of China(61862007);The National Natural Science Foundation of China(62066005);Guangxi Natural Science Foundation(2020GXNSFBA297103);Scientific Research Ability Improving Foundation for Young and Middle-Aged University Teachers in Guangxi(2020KY04030);The School Introduces Talents to Start Scientific Research Projects(2019KJQD17)

摘要/Abstract

摘要：

为了解决传统基于最优化方法所设计的无线网络资源管理策略通常复杂度较高且实时性差，不利于在线决策制定的问题，针对基于SWIPT的传感云系统，建立汇聚（sink）节点能效最大化问题及其数学模型，然后引入深度学习方法，通过对最优化算法的学习实现更低复杂度与更高实时性的算法设计。为了实现深度学习算法在网络资源分配中的应用，首先将 sink 节点最优能效模型转化为高维可求解形式，设计具有迭代形式的SWIPT-WMMSE算法实现最优波束成形矢量的求解，同时证明了算法的收敛性。然后基于DNN逼近误差的传递过程推导了DNN设计准则，并通过对DNN的训练实现其对SWIPT-WMMSE算法的逼近。最后通过仿真实验分别验证了SWIPT-WMMSE算法与DNN算法的有效性，及DNN算法的逼近效果和在提升系统性能方面的优势。

关键词: 深度学习, 无线携能通信, 汇聚节点, 能效, 波束成形, 深度神经网络

Abstract:

To solve the problems of high complexity and poor real-time performance caused by traditional wireless resource management based on optimization methods, the energy efficiency maximization problem of sink node and its mathematical model were established for SWIPT-enabled sensor-cloud system, then the deep learning method was introduced to realize the solving and online decision-making with lower complexity and higher real-time performance.The mathematical model was transformed into a high-dimensional solvable form, and then a SWIFT-WMMSE algorithm with iterated forms was designed to solve optimal beamforming vector.The convergence of SWIPT-WMMSE algorithm was proved.Then, based on error propagation of DNN approximation, the scale design criteria for the DNN was deduced, and the approximation was realized through DNN training.Finally, the simulation results verify the effectiveness of SWIPT-WMMSE and DNN algorithm, as well as the approximation effect of DNN and its system performance gains.

Key words: deep learning, SWIPT, sink node, energy efficiency, beamforming, deep natural network

中图分类号:

TN92

王哲, 李陶深, 葛丽娜, 张桂芬, 吴敏. 基于深度学习的传感云sink节点最优能效SWIPT波束成形设计[J]. 通信学报, 2021, 42(7): 176-188.

Zhe WANG, Taoshen LI, Lina GE, Guifen ZHANG, Min WU. Optimal energy-efficiency beamforming design for SWIPT-enabled sink in sensor cloud based on deep learning[J]. Journal on Communications, 2021, 42(7): 176-188.

图/表 10

图1

图2

图3

图4

表1

图5

图6

图7

表2

表3

参考文献 37

[1]	ALAMRI A , ANSARI W S , HASSAN M M ,et al. A survey on sensor-cloud:architecture,applications,and approaches[J]. International Journal of Distributed Sensor Networks, 2013,9(2): 1-18.
[2]	王田, 沈雪微, 罗皓 ,等. 基于雾计算的可信传感云研究进展[J]. 通信学报, 2019,40(3): 170-181.
	WANG T , SHEN X W , LUO H ,et al. Research progress of trusted sensor-cloud based on fog computing[J]. Journal on Communications, 2019,40(3): 170-181.
[3]	王哲, 李陶深, 葛丽娜 ,等. 基于无线携能通信的传感云系统 sink节点最优能效策略[J]. 控制与决策,2019:doi.org/10.13195/j.kzyjc.2019.1628.
	WANG Z , LI T S , GE L N ,et al. Optimal energy efficiency optimiza-tion strategy for SWIPT-enabled sensor-cloud System[J]. Control and Decision,2019:doi.org/10.13195/j.kzyjc.2019.1628.
[4]	CLERCKX B , ZHANG R , SCHOBER R ,et al. Fundamentals of wireless information and power transfer:from RF energy harvester models to signal and system designs[J]. IEEE Journal on Selected Areas in Communications, 2019,37(1): 4-33.
[5]	王哲, 李陶深, 叶进 ,等. 基于非线性能量收集的多用户MIMO认知无线供电通讯网络[J]. 控制与决策, 2020,35(3): 547-558.
	WANG Z , LI T S , YE J ,et al. Multi-user MIMO cognitive wireless powered communication network based on non-linear energy harvest-ing[J]. Control and Decision, 2020,35(3): 547-558.
[6]	GREGOR K , LECUN Y . Learning fast approximations of sparse coding[C]// Proceedings of the 27th International Conference on Machine Learning. Piscataway:IEEE Press, 2010: 399-406.
[7]	HERSHEY J R , ROUX J L , WENINGER F . Deep unfolding:model-based inspiration of novel deep architectures[J]. arXiv Preprint,arXiv:1409.2574, 2014.
[8]	SPRECHMANN P , LITMAN R , YAKAR T B ,et al. Supervised sparse analysis and synthesis operators[C]// Advances in Neural Information Processing Systems. Massachusetts:MIT Press, 2013: 908-916.
[9]	SUN H , CHEN X , SHI Q ,et al. Learning to optimize:training deep neural networks for interference management[C]// IEEE 18th International Workshop on Signal Processing Advances in Wireless Communications. Piscataway:IEEE Press, 2017: 1-6.
[10]	MEHRABI M , MOHAMMADKARIMI M , ARDAKANI M ,et al. Decision directed channel estimation based on deep neural network$k$-step predictor for MIMO communications in 5G[J]. IEEE Journal on Selected Areas in Communications, 2019,37(11): 2443-2456.
[11]	CUI W , SHEN K M , YU W . Spatial deep learning for wireless scheduling[J]. IEEE Journal on Selected Areas in Communications, 2019,37(6): 1248-1261.
[12]	HE C F , HU Y , CHEN Y ,et al. Joint power allocation and channel assignment for NOMA with deep reinforcement learning[J]. IEEE Journal on Selected Areas in Communications, 2019,37(10): 2200-2210.
[13]	ORTIZ A , ASADI A , ENGELHARDT M ,et al. CBMoS:combinatorial bandit learning for mode selection and resource allocation in D2D systems[J]. IEEE Journal on Selected Areas in Communications, 2019,37(10): 2225-2238.
[14]	NASIR Y S , GUO D N . Multi-agent deep reinforcement learning for dynamic power allocation in wireless networks[J]. IEEE Journal on Selected Areas in Communications, 2019,37(10): 2239-2250.
[15]	GARCIA V , ZHOU Y Q , SHI J L . Coordinated multipoint transmission in dense cellular networks with user-centric adaptive clustering[J]. IEEE Transactions on Wireless Communications, 2014,13(8): 4297-4308.
[16]	LIU L , ZHOU Y Q , ZHUANG W H ,et al. Tractable coverage analysis for hexagonal macrocell-based heterogeneous UDNs with adaptive interference-aware CoMP[J]. IEEE Transactions on Wireless Communications, 2019,18(1): 503-517.
[17]	LIU L , ZHOU Y Q , GARCIA V ,et al. Load aware joint CoMP clustering and inter-cell resource scheduling in heterogeneous ultra dense cellular networks[J]. IEEE Transactions on Vehicular Technology, 2018,67(3): 2741-2755.
[18]	JU H , ZHANG R . Throughput maximization in wireless powered communication networks[J]. IEEE Transactions on Wireless Communications, 2014,13(1): 418-428.
[19]	GONG S Q , MA S D , XING C W ,et al. Optimal beamforming and time allocation for partially wireless powered sensor networks with downlink SWIPT[J]. IEEE Transactions on Signal Processing, 2019,67(12): 3197-3212.
[20]	LIU T , WANG X D , ZHENG L . A cooperative SWIPT scheme for wirelessly powered sensor networks[J]. IEEE Transactions on Communications, 2017,65(6): 2740-2752.
[21]	ZHOU Y Q , LIU L , WANG L ,et al. Service-aware 6G:an intelligent and open network based on the convergence of communication,computing and caching[J]. Digital Communications and Networks, 2020,6(3): 253-260.
[22]	ZHOU Y Q , TIAN L , LIU L ,et al. Fog computing enabled future mobile communication networks:a convergence of communication and computing[J]. IEEE Communications Magazine, 2019,57(5): 20-27.
[23]	QI Y L , TIAN L , ZHOU Y Q ,et al. Mobile edge computing-assisted admission control in vehicular networks:the convergence of communication and computation[J]. IEEE Vehicular Technology Magazine, 2019,14(1): 37-44.
[24]	LYU X C , REN C S , NI W ,et al. Optimal online data partitioning for geo-distributed machine learning in edge of wireless networks[J]. IEEE Journal on Selected Areas in Communications, 2019,37(10): 2393-2406.
[25]	WANG S Q , TUOR T , SALONIDIS T ,et al. Adaptive federated learning in resource constrained edge computing systems[J]. IEEE Journal on Selected Areas in Communications, 2019,37(6): 1205-1221.
[26]	ZHANG R , HO C K . MIMO broadcasting for simultaneous wireless information and power transfer[J]. IEEE Transactions on Wireless Communications, 2013,12(5): 1989-2001.
[27]	WIN M Z , WINTERS J H . Analysis of hybrid selection/maximal-ratio combining in Rayleigh fading[J]. IEEE Transactions on Communications, 1999,47(12): 1773-1776.
[28]	王正强, 蒋铃鸽, 何晨 . 基于合作博弈的多信道认知无线网络中的频谱共享算法[J]. 通信学报, 2014,35(2): 70-75.
	WANG Z Q , JIANG L G , HE C . Spectrum sharing algorithm in mul-ti-channel cognitive radio network based on cooperative game theo-retic[J]. Journal on Communications, 2014,35(2): 70-75.
[29]	SHI Q J , RAZAVIYAYN M , LUO Z Q ,et al. An iteratively weighted MMSE approach to distributed sum-utility maximization for a MIMO interfering broadcast channel[J]. IEEE Transactions on Signal Processing, 2011,59(9): 4331-4340.
[30]	王哲, 李陶深, 叶进 ,等. 基于不确定理论的能量收集可靠性建模及规划[J]. 通信学报, 2018,39(5): 166-176.
	WANG Z , LI T S , YE J ,et al. Reliability modeling and planning of energy harvesting based on uncertainty theory[J]. Journal on Commu-nications, 2018,39(5): 166-176.
[31]	VERDU S . Multiuser detection[M]. Cambridge: Cambridge University Press, 1998.
[32]	SOLODOV M V . On the convergence of constrained parallel variable distribution algorithms[J]. SIAM Journal on Optimization, 1998,8(1): 187-196.
[33]	KUMAR N S , KUMAR K . Performance analysis of m*n equalizer based minimum mean square error (MMSE) receiver for MIMO wireless channel[J]. International Journal of Computer Applications, 2011,6(9): 47-50.
[34]	SUSSMANN H J . Optimal control[M]. Berlin: Springer, 1998.
[35]	HORNIK K , STINCHCOMBE M , WHITE H . Multilayer feed forward networks are universal approximators[J]. Neural Networks, 1989,2(5): 359-366.
[36]	KONE?NY J , LIU J , RICHTáRIK P ,et al. Mini-batch semi-stochastic gradient descent in the proximal setting[J]. IEEE Journal of Selected Topics in Signal Processing, 2015,10(2): 242-255.
[37]	LIAO W C , HONG M Y , LIU Y F ,et al. Base station activation and linear transceiver design for optimal resource management in heterogeneous networks[J]. IEEE Transactions on Signal Processing, 2014,62(15): 3939-3952.

参数	值
小区内传感器节点数K	10
sink节点天线数M_i	5
sink节点发送功率P_i/W	50
数据生成样本总数	30 000
DNN训练样本数	25 000
DNN测试样本数	5 000
批大小bs	10, 50, 100, 500, 1 000
学习率lr	0.01, 0.05, 0.1, 0.5

模型参数	训练阶段参数	最大系统能效/(bit.(Hz.J)^-1)		拟合度：DNN/SWIPT-WMMSE			总CPU用时/s
模型参数	训练阶段参数	DNN	SWIPT-WMMSE	最高	最低	平均	DNN	SWIPT-WMMSE
K=10，M_i=5		7 745	7 801	99.28%	86.09%	94.77%	7.39	255.90
K=20，M_i=5	lr=0.1，bs=10	7 464	7 857	99.00%	82.45%	94.60%	7.62	617.38
K=50，M_i=5		6 852	7 135	98.04%	72.97%	89.80%	8.15	1004.94
K=10，M_i=2	lr=0.1，bs=10	7 532	7 873	98.67%	88.29%	95.51%	7.12	237.22
K=10，M_i=8		7 121	7 516	98.74%	87.60%	92.88%	7.47	319.92
K=10，M_i=5	lr=0.01，bs=10	7 555	7 904	98.58%	73.31%	79.89%	7.41	268.34
	lr=0.05，bs=10	7 460	7 762	99.11%	79.13%	86.79%	7.27	257.92
K=10，M_i=5	lr=0.1，bs=100	7 184	7 283	98.64%	67.70%	86.32%	6.53	248.83
	lr=0.1，bs=1 000	6 348	6 850	92.67%	52.40%	80.91%	6.08	274.74

隐藏层数量		准确度（DNN/SWIPT-WMMSE）			CPU用时（DNN）/s
层	神经元	最高	最低	平均	CPU用时（DNN）/s
3	(1 000,100,10)	99.28%	86.09%	94.77%	7.39
3	(2 000,500,500)	73.23%	7.76%	28.75%	8.89
3	(2 000,1 000,1 000)	71.29%	7.57%	31.75%	10.39
5	(1 000,1 000,500,100,100)	55.37%	3.68%	26.65%	13.83
10	(1 000,1 000,800,800,500, 500,500,300,300,100)	51.17%	4.48%	26.98%	17.35

基于深度学习的传感云sink节点最优能效SWIPT波束成形设计

Optimal energy-efficiency beamforming design for SWIPT-enabled sink in sensor cloud based on deep learning

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 37

相关文章 15

Metrics

推荐阅读 0

[1]	陈东昱, 陈华, 范丽敏, 付一方, 王舰. 基于深度学习的随机性检验策略研究[J]. 通信学报, 2023, 44(6): 23-33.
[2]	李荣鹏, 汪丙炎, 张宏纲, 赵志峰. 知识增强的语义通信接收端设计[J]. 通信学报, 2023, 44(6): 70-76.
[3]	马帅, 裴科, 祁华艳, 李航, 曹雯, 王洪梅, 熊海良, 李世银. 基于生成模型的地磁室内高精度定位算法研究[J]. 通信学报, 2023, 44(6): 211-222.
[4]	林云, 徐怀韬, 王森, 张思成, 庄龙. 基于特征融合的通信语音干扰效果客观评估[J]. 通信学报, 2023, 44(3): 105-116.
[5]	张雷, 王玉, 田建杰, 张琳, 章天骄. 基于IRS辅助的MIMO车联网系统联合波束成形设计[J]. 通信学报, 2023, 44(2): 59-69.
[6]	杨正, 郑云, 余月好, 吴怡, 董志诚, 邢松. 基于自适应功率分裂的协作非正交多址接入无线携能通信网络性能分析[J]. 通信学报, 2023, 44(1): 177-188.
[7]	曾嵘, 杭潇. 车联网环境下可重构智能反射面辅助无线信道估计算法[J]. 通信学报, 2022, 43(8): 142-150.
[8]	郭海燕, 杨震, 邹玉龙, 吕斌, 冯蕴天, 赵玉娟. 基于主被动波束成形联合优化的双RIS辅助抗干扰通信方法[J]. 通信学报, 2022, 43(7): 21-30.
[9]	何世文, 袁军, 安振宇, 张敏, 黄永明, 张尧学. 基于图神经网络的联合用户调度与波束成形优化算法[J]. 通信学报, 2022, 43(7): 73-84.
[10]	杨洁, 董标, 付雪, 王禹, 桂冠. 基于轻量化分布式学习的自动调制分类方法[J]. 通信学报, 2022, 43(7): 134-142.
[11]	杨秀璋, 彭国军, 李子川, 吕杨琦, 刘思德, 李晨光. 基于Bert和BiLSTM-CRF的APT攻击实体识别及对齐研究[J]. 通信学报, 2022, 43(6): 58-70.
[12]	廖勇, 王世义. 高速移动环境下基于RM-Net的大规模MIMO CSI反馈算法[J]. 通信学报, 2022, 43(5): 166-176.
[13]	廖育荣, 王海宁, 林存宝, 李阳, 方宇强, 倪淑燕. 基于深度学习的光学遥感图像目标检测研究进展[J]. 通信学报, 2022, 43(5): 190-203.
[14]	赵增华, 童跃凡, 崔佳洋. 基于域自适应的Wi-Fi指纹设备无关室内定位模型[J]. 通信学报, 2022, 43(4): 143-153.
[15]	李中捷, 熊吉源, 高伟, 韦金迎. 分布式IRS辅助毫米波MU-MISO系统联合波束成形设计[J]. 通信学报, 2022, 43(4): 216-226.