基于时序-空间特征自适应融合的非正交PSWFs信号检测方法

陈文华; 毛忠阳; 陆发平; 孙烨; 高翊轩

doi:10.11999/JEIT260024

基于时序-空间特征自适应融合的非正交PSWFs信号检测方法

doi: 10.11999/JEIT260024 cstr: 32379.14.JEIT260024

陈文华^{1, 2, 3},
毛忠阳^{1, 2},
陆发平^{1, 2, ,},
孙烨³,
高翊轩⁴

1.
中国人民解放军海军航空大学烟台 264001
2.
山东省海空信息感知与处理技术重点实验室烟台 264001
3.
中国人民解放军91428部队宁波 315456
4.
哈尔滨工程大学哈尔滨 150001

基金项目: 博士后创新人才支持计划(BX20200039)，山东省自然科学基金(ZR2023MD045)

详细信息

作者简介:
陈文华：男，博士研究生，工程师，研究方向为现代通信理论与应用、通信信号波形设计、非正弦通信

毛忠阳：男，教授，研究方向为现代通信理论与应用、激光通信、无线通信组网技术

陆发平：男，讲师，研究方向为现代通信理论与应用、测控与通信信号波形设计、非正弦通信

孙烨：男，工程师，研究方向为现代通信理论与应用、通信对抗技术

高翊轩：男，硕士研究生，研究方向为现代通信理论与应用、深度学习

通讯作者:
陆发平　lufaping@163.com

中图分类号: TN911.3
计量
- 文章访问数: 27
- HTML全文浏览量: 16
- PDF下载量: 2
- 被引次数: 0
出版历程
- 修回日期: 2026-07-21
- 录用日期: 2026-07-21
- 网络出版日期: 2026-07-30

Non-Orthogonal PSWFs Signal Detection Method Based on Adaptive Time-Space Feature Fusion

CHEN Wenhua^{1, 2, 3},
MAO Zhongyang^{1, 2},
LU Faping^{1, 2
, ,},
SUN Ye³,
GAO Yixuan⁴

1.
Naval Aviation University, People’s Liberation Army of China, Yantai 264001, China
2.
Shandong Provincial Key Laboratory of Sea-Air Information Perception and Processing Technology, Yantai 264001, China
3.
No. 91428 Force, People’s Liberation Army of China, Ningbo 315456, China
4.
Harbin Engineering University, Harbin 150001, China

Funds: China National Postdoctoral program for Innovative Talents(BX20200039),The National Natural Science Foundation of Shandong Province(ZR2023MD045)

摘要

摘要: 针对非正交椭圆球面波函数(Prolate Spheroidal Wave Functions, PSWFs)信号间相互干扰严重，传统检测方法难以有效分离干扰导致性能下降的难题，提出了一种基于时序-空间特征自适应融合(Adaptive Temporal-Spatial Feature Fusion, ATSFF)的非正交PSWFs信号检测方法。该方法构建时序-空间双路径并行特征提取架构，分别提取信号的一维时序相关特征与二维空间结构特征，设计基于预测置信度的自适应概率加权融合机制实现特征互补增强。仿真结果表明，在误比特率为 4×10^-5时，所提方法系统误码性能比交叉项检测方法提升约0.2 dB。
- 椭圆球面波函数 /
- 非正交信号检测 /
- 时序-空间特征融合 /
- 自适应加权融合
Abstract: Objective To address B5G/6G demands for high spectral efficiency and transmission reliability, non-orthogonal Prolate Spheroidal Wave Function (PSWF) modulation has drawn extensive research interest for its strong time-frequency energy concentration. However, severe mutual interference in PSWF signal multiplexing degrades conventional detection performance in complex channels. Limited by ideal channel assumptions or single-modal feature extraction, existing methods cannot fully exploit signal temporal-spatial information and lack adaptability to dynamic interference. This work proposes an intelligent detection architecture with adaptive temporal-spatial feature fusion (ATSFF) for accurate, robust non-orthogonal PSWF signal detection. Method A dual-path parallel framework extracts complementary features from time and transform domains. A Gated Recurrent Unit (GRU) extracts deep temporal features and captures long-range dependencies from 1D received signals. The other path converts 1D signals into 2D representations via the Gramian Angular Difference Field (GADF), and extracts hierarchical spatial features using ResNet50. An adaptive probability-weighted fusion mechanism dynamically adjusts feature contributions based on sub-network prediction uncertainty to generate robust detection outputs. Results and Discussion Simulations on a 32-class non-orthogonal PSWF dataset (Fig. 2) show the proposed ATSFF outperforms conventional coherent detection, cross-term suppression detection, Approximate Message Passing Interleave Division Multiple Access (AMP-IDMA) and Temporal Sparse Bayesian Learning Least Squares (TMSBL-LS) across the full Signal-to-Noise Ratio (SNR) range. t-SNE visualization (Fig. 4) confirms better inter-class separability and intra-class compactness of fused features. At a bit error rate of 4×10^–5, it achieves 0.2 dB gain (Fig. 6) over coherent detection. ATSFF has higher network overhead and weaker real-time performance than linear detection, yet it supports GPU batch inference with fixed single-sample computation cost, suiting accuracy-sensitive communication scenarios. Conclusions Targeting interference issues in non-orthogonal PSWF signal detection, this paper presents an adaptive temporal-spatial feature fusion detection method. It realizes dual-modal feature extraction via GRU and ResNet50, and adopts a prediction-uncertainty-based adaptive weighted fusion mechanism to integrate dual-path strengths, significantly improving detection accuracy and robustness. Simulations validate its superior performance and stable channel adaptability, providing an efficient solution and design reference for intelligent detection of non-orthogonal PSWF signals.
- prolate spheroidal wave functions /
- non-orthogonal signal detection /
- time-space feature fusion /
- adaptive weighted fusion

HTML全文

图 1 ATSFF架构

下载: 全尺寸图片幻灯片

图 2 GADF转换前后图像

下载: 全尺寸图片幻灯片

图 3 GRU单元结构

下载: 全尺寸图片幻灯片

图 4 t-SNE散点变化

下载: 全尺寸图片幻灯片

图 5 不同比特信噪比误码性能分析

下载: 全尺寸图片幻灯片

图 6 不同信号检测方法系统误码性能

下载: 全尺寸图片幻灯片

表 1 模型复杂度

模型	时间复杂度
相干检测	$ O({N}^{2}) $
AMP-IDMA	$ O(K{N}_{t}{N}_{r}+K{N}_{t}M) $
TMSBL-LS	$ O({K}^{3}),K\gg N $
交叉项检测	$ O({N}^{3}) $
ATSFF	$ O\left(\displaystyle\sum \nolimits_{l=1}^{49}{F}_{l}\cdot {C}_{l}\cdot K_{l}^{2}\cdot {H}_{l}\cdot {W}_{l}\right) $

下载: 导出CSV

表 2 模型超参数

参数类别	参数名称	参数值
训练配置	优化器	Adam
	学习率	0.001
	批次大小	100
	最大训练轮数	200
	早停机制耐心值	15
	权重衰减系数	1e-4
	损失函数	CrossEntropyLoss
网络特定参数	ResNet50权重初始化	ImageNet 预训练权重
网络特定参数	GRU权重初始化	随机初始化

下载: 导出CSV

参考文献(27)

[1]	SADIA H, HASSAN A K, ABBAS Z H, et al. IRS-enabled NOMA communication systems: A network architecture primer with future trends and challenges[J]. Digital Communications and Networks, 2024, 10(5): 1503–1528. doi: 10.1016/j.dcan.2023.09.002.
[2]	CHOWDHURY R and CHOYON A K M S J. Design and performance analysis of spectral-efficient hybrid CPDM-CO-OFDM FSO communication system under diverse weather conditions[J]. Journal of Optical Communications, 2023, 44(s1): s1747–s1766. doi: 10.1515/joc-2021-0113.
[3]	徐常志, 靳一, 李立, 等. 面向6G的星地融合无线传输技术[J]. 电子与信息学报, 2021, 43(1): 28–36. doi: 10.11999/JEIT200363. XU Changzhi, JIN Yi, LI Li, et al. Wireless transmission technology of satellite-terrestrial integration for 6G mobile communication[J]. Journal of Electronics & Information Technology, 2021, 43(1): 28–36. doi: 10.11999/JEIT200363.
[4]	LIU Rui, GUO Kefeng, LI Xingwang, et al. RIS-empowered satellite-aerial-terrestrial networks with PD-NOMA[J]. IEEE Communications Surveys & Tutorials, 2024, 26(4): 2258–2289. doi: 10.1109/COMST.2024.3393612.
[5]	SARKAR D, YOGITA, YADAV S S, et al. A comprehensive survey on IRS-assisted NOMA-based 6G wireless network: Design perspectives, challenges and future directions[J]. IEEE Transactions on Network and Service Management, 2024, 21(2): 2539–2562. doi: 10.1109/TNSM.2023.3348138.
[6]	ZHOU Wenjing, SHEN Mingwei, WU Di, et al. Efficient gridless wideband sparse array synthesis with tapped delay-lines[J]. Digital Signal Processing, 2025, 157: 104893. doi: 10.1016/j.dsp.2024.104893.
[7]	BONAMI A, KERKYACHARIAN G, and PETRUSHEV P. Gaussian bounds for the heat kernel associated to prolate spheroidal wave functions with applications[J]. Constructive Approximation, 2023, 57(2): 351–403. doi: 10.1007/s00365-022-09595-0.
[8]	DENG Kefeng and XUE Rui. Preamble design for frame synchronization in VLF communication based on prolate spheroidal wave function sequences[J]. IEEE Communications Letters, 2025, 29(8): 1948–1952. doi: 10.1109/LCOMM.2025.3581198.
[9]	ALPARONE L, ARIENZO A, and LOMBARDINI F. Improved coherent processing of synthetic aperture radar data through speckle whitening of single-look complex images[J]. Remote Sensing, 2024, 16(16): 2955. doi: 10.3390/rs16162955.
[10]	WANG Guoyang, WU Ruoxi, ZHANG Liangliang, et al. Coherent detection of pulsed terahertz waves in solid, gaseous, and liquid media[J]. Fundamental Research, 2025, 5(2): 516–525. doi: 10.1016/j.fmre.2024.08.003.
[11]	陆发平, 毛忠阳, 许志超, 等. 基于信号间交叉项的非正交椭圆球面波信号检测方法[J]. 电子与信息学报, 2025, 47(8): 2714–2723. doi: 10.11999/JEIT250052. LU Faping, MAO Zhongyang, XU Zhichao, et al. Non-orthogonal prolate spheroidal wave functions signal detection method with cross-terms[J]. Journal of Electronics & Information Technology, 2025, 47(8): 2714–2723. doi: 10.11999/JEIT250052.
[12]	黄永明, 王正. 基于广义投影梯度下降算法的深度学习大规模MIMO信号检测[J]. 东南大学学报: 自然科学版, 2024, 54(4): 961–971. doi: 10.3969/j.issn.1001-0505.2024.04.020. HUANG Yongming and WANG Zheng. Generalizing projected gradient descent algorithm for massive MIMO detection based on deep-learning[J]. Journal of Southeast University: Natural Science Edition, 2024, 54(4): 961–971. doi: 10.3969/j.issn.1001-0505.2024.04.020.
[13]	蒲旭敏, 刘雁翔, 宋米雪, 等. 基于模型驱动深度学习的OTFS信道估计[J]. 电子与信息学报, 2024, 46(2): 680–687. doi: 10.11999/JEIT230072. PU Xumin, LIU Yanxiang, SONG Mixue, et al. Orthogonal time frequency space channel estimation based on model-driven deep learning[J]. Journal of Electronics & Information Technology, 2024, 46(2): 680–687. doi: 10.11999/JEIT230072.
[14]	王骥, 李子龙, 肖健, 等. 基于低复杂度加法网络的非正交多址接入短报文多用户检测算法研究[J]. 电子与信息学报, 2024, 46(6): 2409–2417. doi: 10.11999/JEIT231186. WANG Ji, LI Zilong, XIAO Jian, et al. Research on multi-user detection algorithm for non-orthogonal multiple access short message based on low complexity adder network[J]. Journal of Electronics & Information Technology, 2024, 46(6): 2409–2417. doi: 10.11999/JEIT231186.
[15]	KUMAR A and KUMAR K. Deep learning-based joint NOMA signal detection and power allocation in cognitive radio networks[J]. IEEE Transactions on Cognitive Communications and Networking, 2022, 8(4): 1743–1752. doi: 10.1109/TCCN.2022.3193389.
[16]	KUMAR A, NANTHAAMORNPHONG A, ALSHARIF M H, et al. SIC based RL for massive MIMO NOMA signal detection for different modulation schemes under diverse channel conditions[J]. Scientific Reports, 2025, 15(1): 24641. doi: 10.1038/s41598-025-06492-x.
[17]	WANG Xiaoming, ZHANG Dongcai, CHEN Bingcen, et al. Deep transfer learning for model-driven signal detection in MIMO-NOMA systems[J]. IEEE Transactions on Vehicular Technology, 2023, 72(10): 13039–13054. doi: 10.1109/TVT.2023.3274561.
[18]	李有明, 陈江涛, 吴永宏, 等. 脉冲噪声干扰下免调度非正交多址接入系统的联合多用户检测和信道估计方法[J]. 通信学报, 2025, 46(3): 122–130. doi: 10.11959/j.issn.1000-436x.2025055. LI Youming, CHEN Jiangtao, WU Yonghong, et al. Joint multi-user detection and channel estimation method for grant-free NOMA systems under impulsive noise interference[J]. Journal on Communications, 2025, 46(3): 122–130. doi: 10.11959/j.issn.1000-436x.2025055.
[19]	康洁思, 朱静. 基于交织多址技术的非正交多用户新型检测方法[J]. 电子技术与软件工程, 2021(2): 32–33. doi: 10.20109/j.cnki.etse.2021.02.014. KANG Jiesi and ZHU Jing. A novel non-orthogonal multi-user detection method based on interleave division multiple access[J]. Electronic Technology & Software Engineering, 2021(2): 32–33. doi: 10.20109/j.cnki.etse.2021.02.014.
[20]	XIONG Weihua, JIANG Yufei, and CAO Lixian. Multi-feature fusion recognition method of heart sound signal based on GADF and CNN-ViT[J]. Academic Journal of Computing & Information Science, 2025, 8(4): 1–10. doi: 10.25236/AJCIS.2025.080401.
[21]	FENG Min and XU Juncai. Enhancing non-destructive testing in concrete structures: A GADF-CNN approach for defect detection[J]. Journal of Measurements in Engineering, 2024, 12(2): 333–352. doi: 10.21595/jme.2024.23829.
[22]	PANDA B and SINGH P. A deep convolutional-GRU-SVM-based hybrid approach for signal detection of uplink NOMA system[J]. Wireless Personal Communications, 2024, 138(2): 861–878. doi: 10.1007/s11277-024-11529-x.
[23]	XU Shaohua, LI Jingjing, LIU Kun, et al. A parallel GRU recurrent network model and its application to multi-channel time-varying signal classification[J]. IEEE Access, 2019, 7: 118739–118748. doi: 10.1109/ACCESS.2019.2936516.
[24]	LIU Fugang, ZHANG Ziwei, and ZHOU Ruolin. Automatic modulation recognition based on CNN and GRU[J]. Tsinghua Science and Technology, 2022, 27(2): 422–431. doi: 10.26599/TST.2020.9010057.
[25]	邢六一, 张育芝, 汪洋, 等. 基于深度学习的水声索引调制正交啁啾复用信号检测[J]. 应用声学, 2026, 45(2): 485–493. doi: 10.11684/j.issn.1000-310X.2026.02.022. XING Liuyi, ZHANG Yuzhi, WANG Yang, et al. Deep learning based detector for underwater acoustic orthogonal chirp division multiplexing with index modulation[J]. Journal of Applied Acoustics, 2026, 45(2): 485–493. doi: 10.11684/j.issn.1000-310X.2026.02.022.
[26]	覃日升, 徐志, 况华, 等. 基于双分辨率S变换和改进的多尺度ResNet模型的电能质量扰动检测方法[J]. 广东电力, 2024, 37(7): 68–77. doi: 10.3969/j.issn.1007-290X.2024.07.008. QIN Risheng, XU Zhi, KUANG Hua, et al. Detection method of power quality disturbances based on double resolution S transform and improved multi-scale ResNet[J]. Guangdong Electric Power, 2024, 37(7): 68–77. doi: 10.3969/j.issn.1007-290X.2024.07.008.
[27]	WANG Bo, SHEN Lei, WANG Huaxia, et al. Direct sequence spread spectrum (DSSS) signal detection based on eigenvalues local binary pattern residual network (EL-ResNet)[J]. Signal, Image and Video Processing, 2024, 18(5): 4741–4751. doi: 10.1007/s11760-024-03110-7.