HRT-Net：基于注意力赋权分层网络的无蜂窝协作感知方法

刘升恒; 颜贺; 李兴康; 徐大专; 王东明; 黄永明

doi:10.12000/JR24227

HRT-Net：基于注意力赋权分层网络的无蜂窝协作感知方法

DOI: 10.12000/JR24227

刘升恒^{1, 2, ,},
颜贺¹,
李兴康¹,
徐大专²,
王东明^{1, 2},
黄永明^{1, 2, ,}

1.
东南大学信息科学与工程学院南京 211189
2.
紫金山实验室南京 211111

基金项目: 国家科技重大专项资助(2024ZD1300200)，东南大学“至善青年学者”支持计划(2242023R40005)

详细信息

作者简介:
刘升恒，博士，副教授，主要研究方向为智能感知与无线通信、雷达信号处理等

颜　贺，硕士生，主要研究方向为雷达通信一体化

李兴康，博士生，主要研究方向为雷达通信一体化、协作感知

徐大专，博士，教授，主要研究方向为感知信息论、雷达信号处理等

王东明，博士，教授，主要研究方向为无线通信信号处理、分布式大规模MIMO等

黄永明，博士，教授，主要研究方向为MIMO无线通信、协作无线通信以及网络智能技术等

通讯作者:
刘升恒 s.liu@seu.edu.cn

黄永明 huangym@seu.edu.cn

责任主编：刘凡 Corresponding Editor: LIU Fan
中图分类号: TN959
计量
- 文章访问数: 27
- HTML全文浏览量: 10
- PDF下载量: 8
- 被引次数: 0
出版历程
- 收稿日期: 2024-11-15

HRT-Net: A Cell-Free Cooperative Sensing Method Based on Attention-Weighted Hierarchical Network

LIU Shengheng^{1, 2
, ,},
YAN He¹,
LI Xingkang¹,
XU Dazhuan²,
WANG Dongming^{1, 2},
HUANG Yongming^{1, 2
, ,}

1.
School of Information Science and Engineering, Southeast University, Nanjing 211189, China
2.
Purple Mountain Laboratories, Nanjing 211111, China

Funds: The National Science and Technology Major Project (2024ZD1300200), The Fundamental Research Funds for the Central Universities (2242023R40005)

More Information

Corresponding author: LIU Shengheng, s.liu@seu.edu.cn; HUANG Yongming, huangym@seu.edu.cn

摘要

摘要: 该文针对雷达通信一体化系统中多站协作感知的问题，提出了一种基于无蜂窝网络架构的智能框架HRT-Net，用于实现准确且资源高效的位置估计。具体而言，该文首先将感知区域划分为多个子区域，并基于深度可分离卷积设计了一个轻量级的区域选择网络，以识别目标所属的子区域，从而减少计算负担并实现广域覆盖。其次，考虑到多站数据差异性的隐式问题，本文设计了一种分通道单维注意力机制，旨在有效聚合多站的感知数据并增强特征的提取和表示能力，从而生成注意力权重图以加权修正原始特征。最后，基于多尺度和多重残差连接设计了一个目标定位网络，该网络能够提取更加全面和深层的特征并实现多级特征融合，进而可靠地将其映射到目标的位置坐标。仿真及实测实验结果表明，相比于现有方法，HRT-Net在较低计算复杂度和存储开销下，能够实现厘米级的目标定位。
- 雷达通信一体化 /
- 多点协作 /
- 无蜂窝 /
- 深度学习 /
- 注意力机制
Abstract: This paper proposes an intelligent framework based on a cell-free network architecture, called HRT-Net. HRT-Net is designed to enhance multi-station collaborative sensing problems for joint radar and communication systems, offering accurate and resource-efficient target location estimation. First, the sensing area is divided into sub-regions and a lightweight region selection network employing depthwise separable convolution; this approach coarsely identifies the target’s sub-region, reducing computational demands and enabling extensive area coverage. To tackle interstation data disparity, we propose a channel-wise unidimensional attention mechanism. This mechanism aggregates multi-station sensing data effectively, enhancing feature extraction and representation by generating attention weight maps that refine the original features. Finally, we design a target localization network featuring multi-scale and multi-residual connections. This network extracts comprehensive, deep features and achieves multi-level feature fusion, allowing for reliable mapping of data to the target coordinates. Extensive simulations and real-world experiments validate the effectiveness and robustness of our scheme. The results show that compared with the existing methods, HRT-Net achieves centimeter-level target localization with low computational complexity and minimal storage overhead.
- Joint radar and communications /
- Multi-node collaboration /
- Cell-free /
- Deep learning /
- Attention mechanism

HTML全文

图 1 CF-mMIMO雷达通信一体化系统示意图

Figure 1. Schematic diagram of CF-mMIMO joint radar and communications system

下载: 全尺寸图片幻灯片

图 2 近场范围半径示意图

Figure 2. Schematic diagram of near-field range radius

下载: 全尺寸图片幻灯片

图 3 HRT-Net的网络架构

Figure 3. Network architecture of HRT-Net

下载: 全尺寸图片幻灯片

图 4 深度可分离卷积

Figure 4. Depthwise separable convolution

下载: 全尺寸图片幻灯片

图 5 不同SNR下区域选择网络的性能对比

Figure 5. Performance comparison of region selection network under different SNRs

下载: 全尺寸图片幻灯片

图 6 不同网格间隔下区域选择网络的性能对比

Figure 6. Performance comparison of region selection network under different grid intervals

下载: 全尺寸图片幻灯片

图 7 不同轻量级卷积单元层数下区域选择网络的性能对比

Figure 7. Performance comparison of region selection network under different layers of lightweight convolution unit

下载: 全尺寸图片幻灯片

图 8 不同消融实验下区域选择网络的性能对比

Figure 8. Overhead comparison of different localization methods

下载: 全尺寸图片幻灯片

图 9 不同定位方法的精度对比

Figure 9. Accuracy comparison of different localization methods

下载: 全尺寸图片幻灯片

图 10 不同定位方法的开销对比

Figure 10. Overhead comparison of different localization methods

下载: 全尺寸图片幻灯片

图 11 不同发送和接收AP数量下HRT-Net的精度对比

Figure 11. Accuracy comparison of HRT-Net under different numbers of transmitting and receiving APs

下载: 全尺寸图片幻灯片

图 12 HRT-Net的合理性验证

Figure 12. Rationality verification of HRT-Net

下载: 全尺寸图片幻灯片

图 13 实际场景与配置

Figure 13. Experimental scenario and configuration

下载: 全尺寸图片幻灯片

图 14 3%的通信资源占用率对频谱效率的影响

Figure 14. Influence of 3% communication resources occupancy on the sum spectral efficiency

下载: 全尺寸图片幻灯片

图 15 各轨迹在不同定位方法下的CDF

Figure 15. CDF of each trajectory under different localization methods

下载: 全尺寸图片幻灯片

图 16 不同定位方法下的轨迹预测

Figure 16. Trajectory prediction under different localization methods

下载: 全尺寸图片幻灯片

1 分通道单维注意力机制的计算流程

1. Calculation process of sub-channel single dimensional attention mechanism

输入：输入特征$ {\tilde {\boldsymbol{Y}}_{{\mathrm{in}}}} $
输出：加权特征$ {\tilde {\boldsymbol{Y}}_{\mathrm{w}}} $
初始化参数：$ {\bf{aw}}_{\max}^{{\mathrm{real}}} $, $ {\bf{aw}}_{\max}^{{\mathrm{imag}}} $, $ {\bf{aw}}_{{\mathrm{avg}}}^{{\mathrm{real}}} $, $ {\bf{aw}}_{{\mathrm{avg}}}^{{\mathrm{imag}}} $, $ {{\boldsymbol{A}}^{{\mathrm{real}}}} $, $ {{\boldsymbol{A}}^{{\mathrm{imag}}}} $, 　$ \tilde {\boldsymbol{Y}}_{\mathrm{w}}^{{\mathrm{real}}} $, $ \tilde {\boldsymbol{Y}}_{\mathrm{w}}^{{\mathrm{imag}}} $
1. for $ {\mathrm{ch}} = 1,2 $ //1和2分别代表实部和虚部
2. 　for $ {\mathrm{tr}} = 1,2, \cdots ,{\mathrm{TR}} $
3. 　　$ {\bf{aw}}_{\max}^{{\mathrm{ch}}}({\mathrm{tr}}) = \max\left( {\tilde {\boldsymbol{Y}}_{{\mathrm{in}}}^{{\mathrm{ch}}}(1:{\mathrm{KL}},{\mathrm{tr}})} \right) $
4. 　　$ {\bf{aw}}_{{\mathrm{avg}}}^{{\mathrm{ch}}}({\mathrm{tr}}) = \dfrac{1}{{{\mathrm{KL}}}}\displaystyle\sum\limits_{{\mathrm{kl}} = 1}^{{\mathrm{KL}}} {\tilde {\boldsymbol{Y}}_{{\mathrm{in}}}^{{\mathrm{ch}}}({\mathrm{kl}},{\mathrm{tr}})} $
5. 　end
6. 　//$ {\mathrm{conv}} $函数和$ {\mathrm{expand}} $函数分别代表标准卷积操作和扩展维　　　度操作
7. 　$ {{\boldsymbol{A}}^{{\mathrm{ch}}}} = {\mathrm{expand}}({\mathrm{sigmoid}}({\mathrm{conv}}({\mathrm{concat}}({\bf{aw}}_{\max }^{{\mathrm{ch}}},{\bf{aw}}_{{\mathrm{avg}}}^{{\mathrm{ch}}})))) $
8. 　$ \tilde {\boldsymbol{Y}}_{\mathrm{w}}^{{\mathrm{ch}}} = {A^{{\mathrm{ch}}}} \odot \tilde {\boldsymbol{Y}}_{{\mathrm{in}}}^{{\mathrm{ch}}} $ //符号$ \odot $代表元素级相乘
9. end
10. $ {\tilde {\boldsymbol{Y}}_{\mathrm{w}}} = {\mathrm{concat}}(\tilde {\boldsymbol{Y}}_{\mathrm{w}}^{{\mathrm{real}}},\tilde {\boldsymbol{Y}}_{\mathrm{w}}^{{\mathrm{imag}}{\mathrm{w}}}) $

下载: 导出CSV

表 1 仿真参数设置

Table 1. Simulation parameter settings

参数	数值	参数	数值
载波频率	30 GHz	天线间距	半波长
子载波间隔	60 kHz	循环前缀长度	64
子载波数	1024	发送AP坐标	$ (4.8,1.2),(5.2,9.2) $
符号数	2	接收AP坐标	$ (2.4,0),(7.2,0),(2.8,10),(7.6,10) $

下载: 导出CSV

表 2 HRT-Net主要网络参数设置

Table 2. Main network parameter settings of HRT-Net

模块名称		操作	超参数
模块名称		操作	核尺寸	步长	填充	输入通道	输出通道
区域选择网络	基本卷积模块	标准卷积	$ (3,3) $	1	2	2	32
	轻量级卷积模块	第1层DW卷积	$ (3,3) $	1	1	2	2
		第2层DW卷积	$ (3,3) $	1	1	32	32
		第3层DW卷积	$ (3,3) $	1	1	32	32
		第4层DW卷积	$ (3,3) $	1	1	64	64
		第5层DW卷积	$ (3,3) $	1	1	64	64
		第6层DW卷积	$ (3,3) $	1	1	128	128
	回归模块	全连接层	\	\	\	128	4
分通道单维注意力机制		标准卷积	$ (1,3) $	1	$ (0,1) $	2	1
目标定位网络	多尺度卷积模块	标准卷积1	$ (3,3) $	1	$ (1,1) $	2	32
		标准卷积2	$ (5,5) $	1	$ (2,2) $	2	32
		标准卷积3	$ (9,9) $	1	$ (4,4) $	2	32
	精细化卷积模块	第1层标准卷积1	$ (3,3) $	2	1	2	32
		第1层标准卷积2	$ (3,3) $	2	1	32	32
		第2层标准卷积1	$ (3,3) $	2	1	32	64
		第2层标准卷积2	$ (3,3) $	2	1	64	64
		第3层标准卷积1	$ (3,3) $	2	1	64	128
		第3层标准卷积2	$ (3,3) $	2	1	128	128
	回归模块	全连接层	\	\	\	128	2

下载: 导出CSV

表 3 实测场景参数设置

Table 3. Experimental scenario parameter settings

参数	数值	参数	数值
发送AP坐标	$ (2.4,1.2) $	轨迹1	$ (2.4,5.0) \to (2.4,9.6) $
接收AP坐标	$ (0,0),(1.2,0),(3,0),(4.8,0) $	轨迹2	$ (0.6,4.1) \to (4.2,4.1) $
\	\	轨迹3	$ (0.6,9.6) \to (4.8,5.4) $

下载: 导出CSV

参考文献(55)

[1]	MAO Weihao, LU Yang, LIU Jingxian, et al. Beamforming design in cell-free massive MIMO integrated sensing and communication systems[C]. GLOBECOM 2023-2023 IEEE Global Communications Conference, Kuala Lumpur, Malaysia, 2023: 546–551. doi: 10.1109/GLOBECOM54140.2023.10437290.
[2]	马丁友, 刘祥, 黄天耀, 等. 雷达通信一体化: 共用波形设计和性能边界[J]. 雷达学报, 2022, 11(2): 198–212. doi: 10.12000/JR21146. MA Dingyou, LIU Xiang, HUANG Tianyao, et al. Joint radar and communications: Shared waveform designs and performance bounds[J]. Journal of Radars, 2022, 11(2): 198–212. doi: 10.12000/JR21146.
[3]	LIU Fan, ZHENG Le, CUI Yuanhao, et al. Seventy years of radar and communications: The road from separation to integration[J]. IEEE Signal Processing Magazine, 2023, 40(5): 106–121. doi: 10.1109/MSP.2023.3272881.
[4]	YOU Xiaohu, HUANG Yongming, LIU Shengheng, et al. Toward 6G $ {\text{TK}}\mu $ extreme connectivity: Architecture, key technologies and experiments[J]. IEEE Wireless Communications, 2023, 30(3): 86–95. doi: 10.1109/MWC.004.2200482.
[5]	LU Shihang, LIU Fan, LI Yunxin, et al. Integrated sensing and communications: Recent advances and ten open challenges[J]. IEEE Internet of Things Journal, 2024, 11(11): 19094–19120. doi: 10.1109/JIOT.2024.3361173.
[6]	尹浩, 黄宇红, 韩林丛, 等. 6G通信–感知–计算融合网络的思考[J]. 中国科学: 信息科学, 2023, 53(9): 1838–1842. doi: 10.1360/SSI-2023-0135. YIN Hao, HUANG Yuhong, HAN Lincong, et al. Thoughts on 6G integrated communication, sensing and computing networks[J]. SCIENTIA SINICA Informationis, 2023, 53(9): 1838–1842. doi: 10.1360/SSI-2023-0135.
[7]	WANG Yunlong, WU Ying, and SHEN Yuan. Cooperative tracking by multi-agent systems using signals of opportunity[J]. IEEE Transactions on Communications, 2020, 68(1): 93–105. doi: 10.1109/TCOMM.2019.2944605.
[8]	XIE Lei, SONG Shenghui, ELDAR Y C, et al. Collaborative sensing in perceptive mobile networks: Opportunities and challenges[J]. IEEE Wireless Communications, 2023, 30(1): 16–23. doi: 10.1109/MWC.005.2200214.
[9]	NGO H Q, ASHIKHMIN A, YANG Hong, et al. Cell-free massive MIMO versus small cells[J]. IEEE Transactions on Wireless Communications, 2017, 16(3): 1834–1850. doi: 10.1109/TWC.2017.2655515.
[10]	MENG Kaitao, MASOUROS C, PETROPULU A P, et al. Cooperative ISAC networks: Opportunities and challenges[J]. IEEE Wireless Communications, 2024, 1–8. doi: 10.1109/MWC.008.2400151.
[11]	WANG Xinhua, ASHIKHMIN A, and WANG Xiaodong. Wirelessly powered cell-free IoT: Analysis and optimization[J]. IEEE Internet of Things Journal, 2020, 7(9): 8384–8396. doi: 10.1109/JIOT.2020.2990378.
[12]	WEI Chen, XU Kui, SHEN Zhexian, et al. Fingerprint-based localization and channel estimation integration for cell-free massive MIMO IoT systems[J]. IEEE Internet of Things Journal, 2022, 9(24): 25237–25252. doi: 10.1109/JIOT.2022.3195899.
[13]	王鼎, 尹洁昕, 高路, 等. 一种协同二维DOA和TDOA观测量的超视距短波辐射源定位新方法[J]. 雷达学报(中英文), 2024, 13(6): 1135–1156. doi: 10.12000/JR24136. WANG Ding, YIN Jiexin, GAO Lu, et al. A novel cooperative positioning method for over-the-horizon shortwave emitter based on two-dimensional direction-of-arrival and time-difference-of-arrival measurements[J]. Journal of Radars, 2024, 13(6): 1135–1156. doi: 10.12000/JR24136.
[14]	XU Sheng, WU Linlong, DOĞANÇAY K, et al. A hybrid approach to optimal TOA-sensor placement with fixed shared sensors for simultaneous multi-target localization[J]. IEEE Transactions on Signal Processing, 2022, 70: 1197–1212. doi: 10.1109/TSP.2022.3152232.
[15]	YUAN Weijie, WU Nan, GUO Qinghua, et al. TOA-based passive localization constructed over factor graphs: A unified framework[J]. IEEE Transactions on Communications, 2019, 67(10): 6952–6965. doi: 10.1109/TCOMM.2019.2930517.
[16]	朱丹, 徐威远, 陈文娟, 等. 基于光波分复用网络的分布式多目标定位系统[J]. 雷达学报, 2019, 8(2): 171–177. doi: 10.12000/JR19028. ZHU Dan, XU Weiyuan, CHEN Wenjuan, et al. Distributed multi-target localization system based on optical wavelength division multiplexing network[J]. Journal of Radars, 2019, 8(2): 171–177. doi: 10.12000/JR19028.
[17]	DAI Zichao, WANG Gang, JIN Xiaoping, et al. Nearly optimal sensor selection for TDOA-based source localization in wireless sensor networks[J]. IEEE Transactions on Vehicular Technology, 2020, 69(10): 12031–12042. doi: 10.1109/TVT.2020.3011118.
[18]	MENG Fan, LIU Shengheng, GAO Songtao, et al. TDoA positioning with data-driven LoS inference in mmWave MIMO communications[J]. Signal Processing, 2024, 220: 109447. doi: 10.1016/j.sigpro.2024.109447.
[19]	LIU Shengheng, LI Xingkang, MAO Zihuan, et al. Model-driven deep neural network for enhanced AoA estimation using 5G gNB[C]. The 38th AAAI Conference on Artificial Intelligence, Vancouver, Canada, 2024: 214–221. doi: 10.1609/aaai.v38i1.27773.
[20]	WANG Shengchu, JIANG Xianbo, and WYMEERSCH H. Cooperative localization in wireless sensor networks with AOA measurements[J]. IEEE Transactions on Wireless Communications, 2022, 21(8): 6760–6773. doi: 10.1109/TWC.2022.3152426.
[21]	MENG Kaitao, HAN K, MASOUROS C, et al. Network-level ISAC: Performance analysis and optimal antenna-to-BS allocation[J]. arXiv: 2410.06365, 2024.
[22]	ZUO Peiliang, PENG Tao, WU Hao, et al. Directional source localization based on RSS-AOA combined measurements[J]. China Communications, 2020, 17(11): 181–193. doi: 10.23919/JCC.2020.11.015.
[23]	ZHANG Zhenkun, REN Hong, PAN Cunhua, et al. Target localization in cooperative ISAC systems: A scheme based on 5G NR OFDM signals[J]. IEEE Transactions on Communications, 2024, 1. doi: 10.1109/TCOMM.2024.3486981.
[24]	YIN Lu, NI Qiang, and DENG Zhongliang. Intelligent multisensor cooperative localization under cooperative redundancy validation[J]. IEEE Transactions on Cybernetics, 2021, 51(4): 2188–2200. doi: 10.1109/TCYB.2019.2900312.
[25]	FIGUEROA M R, BISHOYI P K, and PETROVA M. Optimal weight scheme for fusion-assisted cooperative multi-monostatic object localization in 6G networks[J]. arXiv: 2408.16464, 2024.
[26]	JIA Ruo, XU Kui, XIA Xiaochen, et al. Time-sequential cooperative localization for moving sensor in millimeter-wave cell-free massive MIMO system[J]. IEEE Sensors Journal, 2022, 22(22): 22008–22019. doi: 10.1109/JSEN.2022.3211743.
[27]	WEI Zhiqing, XU Ruizhong, FENG Zhiyong, et al. Symbol-level integrated sensing and communication enabled multiple base stations cooperative sensing[J]. IEEE Transactions on Vehicular Technology, 2024, 73(1): 724–738. doi: 10.1109/TVT.2023.3304856.
[28]	PUCCI L, BACCHIELLI T, and GIORGETTI A. Cooperative maximum likelihood target position estimation for MIMO-ISAC networks[J]. arXiv: 2411.05187, 2024.
[29]	GAO Peng, LIAN Lixiang, YU Jinpei. Cooperative ISAC with direct localization and rate-splitting multiple access communication: A Pareto optimization framework[J]. IEEE Journal on Selected Areas in Communications, 2023, 41(5): 1496–1515. doi: 10.1109/JSAC.2023.3240714.
[30]	JIANG Wangjun, WEI Zhiqing, LI Bin, et al. Improve radar sensing performance of multiple roadside units cooperation via space registration[J]. IEEE Transactions on Vehicular Technology, 2022, 71(10): 10975–10990. doi: 10.1109/TVT.2022.3184182.
[31]	TAGLIAFERRI D, MANZONI M, MIZMIZI M, et al. Cooperative coherent multistatic imaging and phase synchronization in networked sensing[J]. IEEE Journal on Selected Areas in Communications, 2024, 42(10): 2905–2921. doi: 10.1109/JSAC.2024.3414609.
[32]	HUANG Yanru, MAZUELAS S, GE Feng, et al. Indoor localization system with NLOS mitigation based on self-training[J]. IEEE Transactions on Mobile Computing, 2023, 22(7): 3952–3966. doi: 10.1109/TMC.2022.3148338.
[33]	LIU Guanying, LIU An, ZHANG Rui, et al. Angular-domain selective channel tracking and Doppler compensation for high-mobility mmWave massive MIMO[J]. IEEE Transactions on Wireless Communications, 2021, 20(5): 2902–2916. doi: 10.1109/TWC.2020.3045272.
[34]	WYMEERSCH H and SECO-GRANADOS G. Radio localization and sensing—part II: State-of-the-art and challenges[J]. IEEE Communications Letters, 2022, 26(12): 2821–2825. doi: 10.1109/LCOMM.2022.3206846.
[35]	LIU Junchen, WANG Tianyu, LI Yuxiao, et al. A transformer-based signal denoising network for AoA estimation in NLoS environments[J]. IEEE Communications Letters, 2022, 26(10): 2336–2339. doi: 10.1109/LCOMM.2022.3187661.
[36]	HE Hangchuan, TANG Yupeng, and GUO Wensheng. Target localization in ISAC system based on BP neural network[C]. 2023 IEEE International Conference on Image Processing and Computer Applications (ICIPCA), Changchun, China, 2023: 444–450. doi: 10.1109/ICIPCA59209.2023.10257773.
[37]	TEDESCHINI B C, KWON G, NICOLI M, et al. Real-time Bayesian neural networks for 6G cooperative positioning and tracking[J]. IEEE Journal on Selected Areas in Communications, 2024, 42(9): 2322–2338. doi: 10.1109/JSAC.2024.3413950.
[38]	ZENG Fan, LIU Ruoyun, SUN Xiaoyu, et al. Multi-static ISAC based on network-assisted full-duplex cell-free networks: Performance analysis and duplex mode optimization[J]. arXiv: 2406.08268, 2024.
[39]	ADHIKARY A, RAHA A D, QIAO Yu, et al. Holographic MIMO with integrated sensing and communication for energy-efficient cell-free 6G networks[J]. IEEE Internet of Things Journal, 2024, 11(19): 30617–30635. doi: 10.1109/JIOT.2024.3411695.
[40]	郭帅, 陈婷, 王鹏辉, 等. 基于角度引导Transformer融合网络的多站协同目标识别方法[J]. 雷达学报, 2023, 12(3): 516–528. doi: 10.12000/JR23014. GUO Shuai, CHEN Ting, WANG Penghui, et al. Multistation cooperative radar target recognition based on an angle-guided transformer fusion network[J]. Journal of Radars, 2023, 12(3): 516–528. doi: 10.12000/JR23014.
[41]	WOO S, PARK J, LEE J Y, et al. CBAM: Convolutional block attention module[C]. The 15th European Conference on Computer Vision (ECCV), Munich, Germany, 2018: 3–19. doi: 10.1007/978-3-030-01234-2_1.
[42]	TEMIZ M, ALSUSA E, and BAIDAS M W. A dual-function massive MIMO uplink OFDM communication and radar architecture[J]. IEEE Transactions on Cognitive Communications and Networking, 2022, 8(2): 750–762. doi: 10.1109/TCCN.2021.3128599.
[43]	HOWARD A G, ZHU Menglong, CHEN Bo, et al. Mobilenets: Efficient convolutional neural networks for mobile vision applications[J]. arXiv: 1704.04861, 2017.
[44]	ZHANG Haokui, LI Ying, JIANG Yenan, et al. Hyperspectral classification based on lightweight 3-D-CNN with transfer learning[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(8): 5813–5828. doi: 10.1109/TGRS.2019.2902568.
[45]	KUMAR A, SHARMA A, BHARTI V, et al. MobiHisNet: A lightweight CNN in mobile edge computing for histopathological image classification[J]. IEEE Internet of Things Journal, 2021, 8(24): 17778–17789. doi: 10.1109/JIOT.2021.3119520.
[46]	AMMAR H A, ADVE R, SHAHBAZPANAHI S, et al. User-centric cell-free massive MIMO networks: A survey of opportunities, challenges and solutions[J]. IEEE Communications Surveys & Tutorials, 2022, 24(1): 611–652. doi: 10.1109/COMST.2021.3135119.
[47]	LU Zhilin, WANG Jintao, and SONG Jian. Multi-resolution CSI feedback with deep learning in massive MIMO system[C]. ICC 2020-2020 IEEE International Conference on Communications (ICC), Dublin, Ireland, 2020: 1–6. doi: 10.1109/ICC40277.2020.9149229.
[48]	HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al. Deep residual learning for image recognition[C]. 2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, USA, 2016: 770–778. doi: 10.1109/CVPR.2016.90.
[49]	KINGMA D P and BA J L. Adam: A method for stochastic optimization[C]. The 3rd International Conference on Learning Representations, San Diego, USA, 2015.
[50]	SAKHNINI A, DE BAST S, GUENACH M, et al. Near-field coherent radar sensing using a massive MIMO communication testbed[J]. IEEE Transactions on Wireless Communications, 2022, 21(8): 6256–6270. doi: 10.1109/TWC.2022.3148035.
[51]	DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16x16 words: Transformers for image recognition at scale[C]. The 9th International Conference on Learning Representations, 2021.
[52]	LIU Rui, LI Jiamin, LIU Feiyang, et al. Scalable fingerprint-based localization framework for 3D cell-free radio access networks[J]. IEEE Communications Letters, 2024, 28(8): 1840–1844. doi: 10.1109/LCOMM.2024.3414991.
[53]	YIN Jun and VERHELST M. CNN-based robust sound source localization with SRP-PHAT for the extreme edge[J]. ACM Transactions on Embedded Computing Systems, 2023, 22(3): 55. doi: 10.1145/3586996.
[54]	CAO Yang, WANG Pan, ZHENG Kang, et al. Experimental performance evaluation of cell-free massive MIMO systems using COTS RRU with OTA reciprocity calibration and phase synchronization[J]. IEEE Journal on Selected Areas in Communications, 2023, 41(6): 1620–1634. doi: 10.1109/JSAC.2023.3276057.
[55]	WANG Dongming, YOU Xiaohu, HUANG Yongming, et al. Full-spectrum cell-free RAN for 6G systems: System design and experimental results[J]. Science China Information Sciences, 2023, 66(3): 130305. doi: 10.1007/s11432-022-3664-x.