基于角度间隔分离学习的相干DOA估计方法

doi:10.12305/j.issn.1001-506X.2025.10.07

摘要/Abstract

摘要：

针对传统模型驱动算法处理空间临近相干源信号精度下降，以及数据驱动算法需大样本训练的问题，提出基于角度间隔分离学习的高精度波达方向估计方法。该方法利用信号角度间隔稀疏特性，结合分域思想，通过空间滤波器估计角度间隔信息，将信号划分至对应角度间隔域，再利用深层神经网络多标签分类器完成波达方向估计。同时引入稀疏自编码技术，通过压缩输入数据和提取关键特征，降低计算复杂度的同时有效滤除干扰。仿真结果表明，与其他数据驱动算法相比，该方法在少量训练样本条件下对空间临近角度信号具有更高的估计精度和泛化能力。

关键词: 相干波达方向估计, 角度间隔分离学习, 深层神经网络, 稀疏自编码器

Abstract:

In face of the accuracy degradation of traditional model-driven algorithms when processing spatially close coherent source signals and the large training sample requirement of data-driven algorithms, a high-precision direction of arrival （DOA） estimation method based on angle interval separation learning （AISL） is proposed. It leverages the sparsity of signal angle intervals and employs spatial filters to estimate angle interval information using the concept of area separation. The signals are then partitioned into corresponding angle interval areas, followed by DOA estimation through deep neural network （DNN） multi-label classifiers. Additionally, sparse autoencoder （SAE） technology is introduced to compress input data and extract key features, effectively reducing computational complexity while filtering out interference. Simulation results demonstrate that compared to other data-driven algorithms, this method achieves superior estimation accuracy and generalization ability in spatially close angle domains under limited training sample conditions.

Key words: coherent direction of arrival （DOA） estimation, angular interval separation learning （AISL）, deep neural network （DNN）, sparse autoencoder （SAE）

中图分类号:

TN 911.7

王军, 武子涵, 周广佼, 周志权. 基于角度间隔分离学习的相干DOA估计方法[J]. 系统工程与电子技术, 2025, 47(10): 3188-3198.

Jun WANG, Zihan WU, Guangjiao ZHOU, Zhiquan ZHOU. Coherent DOA estimation method based on angle interval separation learning[J]. Systems Engineering and Electronics, 2025, 47(10): 3188-3198.

图/表 17

图1

图2

图3

表1

表2

图4

图5

图6

图7

图8

图9

图10

图11

图12

图13

图14

图15

参考文献 39

1	SCHMIDT R. Multiple emitter location and signal parameter estimation[J]. IEEE Trans. on Antennas and Propagation, 1986, 34 (3): 276- 280. doi: 10.1109/TAP.1986.1143830
2	HAYKIN S, GREENLAY T, LITVA J. Performance evaluation of the modified FBLP method for angle of arrival estimation using real radar multipath data[J]. Communications, Radar and Signal Processing, 1985, 132 (3): 159- 174.
3	SHAN T J, WAX M, KAILATH T. On spatial smoothing for direction-of-arrival estimation of coherent signals[J]. IEEE Trans. on Acoustics, Speech, and Signal Processing, 1985, 33 (4): 806- 811. doi: 10.1109/TASSP.1985.1164649
4	RAO B D, HARI K V S. Weighted subspace methods and spatial smoothing: analysis and comparison[J]. IEEE Trans. on Signal Processing, 1993, 41 (2): 788- 803. doi: 10.1109/78.193218
5	ROY R, KAILATH T. ESPRIT-estimation of signal parameters via rotational invariance techniques[J]. IEEE Trans. on Acoustics, Speech, and Signal Processing, 1989, 37 (7): 984- 995. doi: 10.1109/29.32276
6	KUMAR A A, ZHANG D Y, LIN Z P, et al. Direction finding and multipath mitigation using single antenna[C]//Proc. of the IEEE International Symposium on Circuits and Systems, 2018.
7	WANG Y L, CHEN H, PENG Y N, et al. Spatial spectrum estimation theory and algorithm[M]. Beijing: Tsinghua University Press, 2004.
8	PAN J J, SUN M, WANG Y D, et al. Simplified spatial smoothing for DOA estimation of coherent signals[J]. IEEE Trans. on Circuits and Systems II: Express Briefs, 2022, 70 (2): 841- 845.
9	ZHAGYPAR R, ZHAGYPAROVA K, AKHTAR M T. Spatially smoothed TF-root-MUSIC for DOA estimation of coherent and non-stationary sources under noisy conditions[J]. IEEE Access, 2021, 9, 95754- 95766. doi: 10.1109/ACCESS.2021.3095345
10	YOU C Y, ZHANG D H. Research on DOA estimation based on spatial smoothing algorithm and optimal smoothing times[C]//Proc. of the IEEE International Conference on Data Science and Computer Application, 2021: 824−829.
11	TIAN G G, ZHAO C L, LI J X, et al. The DOA estimation method based on improved SSMUSIC algorithm[C]//Proc. of the 3rd China International SAR Symposium, 2022.
12	WANG K D, SHI L, CHEN T. Two-dimensional separable gridless direction-of-arrival estimation based on finite rate of innovation[J]. IEEE Access, 2021, 9, 17275- 17283. doi: 10.1109/ACCESS.2021.3054660
13	QI B B.. DOA estimation of the coherent signals using beamspace matrix reconstruction[J]. Signal Processing, 2022, 191, 108349. doi: 10.1016/j.sigpro.2021.108349
14	高世伟, 保铮. 利用数据矩阵分解实现对空间相关信号源的超分辨处理[J]. 通信学报, 1988, 9 (1): 4- 13.
	GAO S W, BAO Z. Using data matrix decomposition to achieve super resolution processing of spatial correlation signal sources[J]. Journal of Communications, 1988, 9 (1): 4- 13.
15	YOU Z Y, HU G P, ZHOU H M, er al. Joint estimation method of DOD and DOA of bistatic coprime array MIMO radar for coherent targets based on low-rank matrix reconstruction[J]. Sensors, 2022, 22 (12): 4625. doi: 10.3390/s22124625
16	VIBERG M, OTTERSTEN B, KAILATH T. Detection and estimation in sensor arrays using weighted subspace fitting[J]. IEEE Trans. on Signal Processing, 1991, 39 (11): 2436- 2449. doi: 10.1109/78.97999
17	CHEN Q, LIU R L. On the explanation of spatial smoothing in MUSIC algorithm for coherent sources[C]//Proc. of the International Conference on Information Science and Technology, 2011: 699−702.
18	王鸿帧, 郑桂妹, 陈晨, 等. 米波雷达低仰角估计技术分析与展望[J]. 火力与指挥控制, 2023, 48 (11): 6- 16. doi: 10.3969/j.issn.1002-0640.2023.11.002
	WANG H Z, ZHENG G M, CHEN C, et al. Analysis and prospect of VHF radar low elevation angle estimation technology[J]. Fire Control & Command Control, 2023, 48 (11): 6- 16. doi: 10.3969/j.issn.1002-0640.2023.11.002
19	RAIGURU P, ROUT S K, SAHANI M, et al. Machine learning aided sparse direction of arrival estimation[J]. IEEE Sensors Journal, 2024, 24 (22): 38125- 38134. doi: 10.1109/JSEN.2024.3453996
20	PAVEL S R, CHOWDHURY M W T S, ZHANG Y D, et al. Machine learning-based direction-of-arrival estimation exploiting distributed sparse arrays[C]//Proc. of the 55th Asilomar Conference on Signals, Systems, and Computers, 2021: 241−245.
21	MASSIMO D, VIANI F, ROCCA P, et al. An innovative multiresolution approach for DOA estimation based on a support vector classification[J]. IEEE Trans. on Antennas and Propagation, 2009, 57 (8): 2279- 2292. doi: 10.1109/TAP.2009.2024485
22	RAHMAN S, MANDALAM Y, ADHIKARI K. Performance analysis of SVM-based DOA estimation for uniform linear arrays[C]//Proc. of the IEEE 14th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference, 2023: 339−345.
23	AL KASSIR H, KANTARTZIS N V, LAZARIDIS P I, et al. Improving DOA Estimation via an optimal deep residual neural network classifier on uniform linear arrays[J]. IEEE Open Journal of Antennas and Propagation, 2024, 5 (2): 460- 473. doi: 10.1109/OJAP.2024.3362061
24	WU Z H, WANG J, ZHOU Z Q. Two-dimensional coherent polarization-direction-of-arrival estimation based on sequence-embedding fusion transformer[J]. Remote Sensing, 2024, 16 (21): 3977. doi: 10.3390/rs16213977
25	GUO Y, ZHANG Z, HUANG Y H. Dual class token vision transformer for direction of arrival estimation in low SNR[J]. IEEE Signal Processing Letters, 2023, 31, 76- 80.
26	ZHANG Y H, ZENG R, ZHANG S C, et al. Complex-valued neural network with multi-step training for single-snapshot DOA estimation[J]. IEEE Geoscience and Remote Sensing Letters, 2024, 21, 1- 5.
27	刘其有, 何瑞, 施伟. 基于深度学习的波达方向估计方法综述[J]. 中国电子科学研究院学报, 2025, 20(1): 1−9.
	LIU Q Y, HE R, SHI W. A survey on direction of arrival estimation methods based on deep learning[J]. Journal of China Academy of Electronics and Information Technology, 2025, 20(1): 1−9.
28	王鸿帧, 郑桂妹, 宋玉伟. 基于深度学习的米波雷达低仰角估计方法[J]. 现代雷达, 2022, 44 (12): 15- 24.
	WANG H Z, ZHENG G M, SONG Y W. Deep learning-based VHF radar low elevation angle estimation method[J]. Modern Radar, 2022, 44 (12): 15- 24.
29	WU L, LIU Z M, HUANG Z T. Deep convolution network for direction of arrival estimation with sparse prior[J]. IEEE Signal Process Letters, 2019, 26 (11): 1688- 1692. doi: 10.1109/LSP.2019.2945115
30	QIN Y H. Deep networks for direction of arrival estimation with sparse prior in low SNR[J]. IEEE Access, 2023, 11, 44637- 44648. doi: 10.1109/ACCESS.2023.3273126
31	MERKOFER J P, REVACH G, SHLEZINGER N, et al. DA-MUSIC: data-driven DoA estimation via deep augmented MUSIC algorithm[J]. IEEE Trans. on Vehicular Technology, 2023, 73 (2): 2771- 2785.
32	CHEN P, CHEN Z M, LIU L, et al. SDOA-Net: an efficient deep learning-based DOA estimation network for imperfect array[J]. IEEE Trans. on Instrumentation and Measurement, 2024, 73
33	CHEN D W, SHI S, GU X M, et al. Robust DoA estimation using denoising autoencoder and deep neural networks[J]. IEEE Access, 2022, 10, 52551- 52564. doi: 10.1109/ACCESS.2022.3164897
34	YAO Y Y, LEI H, HE W J. A-CRNN-based method for coherent DOA estimation with unknown source number[J]. Sensors, 2020, 20 (8): 2296. doi: 10.3390/s20082296
35	LIU Z M, ZHANG C, YU P S. Direction-of-arrival estimation based on deep neural networks with robustness to array imperfections[J]. IEEE Trans. on Antennas and Propagation, 2018, 66 (12): 7315- 7327. doi: 10.1109/TAP.2018.2874430
36	XIANG H H, CHEN B X, YANG T, et al. Phase enhancement model based on supervised convolutional neural network for coherent DOA estimation[J]. Applied Intelligence, 2020, 50, 2411- 2422. doi: 10.1007/s10489-020-01678-4
37	XIANG H H, CHEN B X, YANG M L, et al. Angle separation learning for coherent DOA estimation with deep sparse prior[J]. IEEE Communications Letters, 2021, 25 (2): 465- 469. doi: 10.1109/LCOMM.2020.3032733
38	SAVARGAONKAR M, OYEWOLE I, CHEHADE A, et al. Uncorrelated sparse autoencoder with long short-term memory for state-of-charge estimations in lithium-ion battery cells[J]. IEEE Trans. on Automation Science and Engineering, 2024, 21 (1): 15- 26. doi: 10.1109/TASE.2022.3222759
39	刘恺忻, 付进, 邹男, 等. 利用协方差矩阵拟合的阵列孔径扩展方法[J]. 声学学报, 2023, 48 (5): 911- 919.
	LIU K X, FU J, ZOU N, et al. Array aperture extension method using covariance matrix fitting[J]. Acta Acustica, 2023, 48 (5): 911- 919.

参数	取值	参数	取值
阵元数M	$21$	学习率$\alpha $	$ 0.002$
波长$\lambda /{\mathrm{m}} $	$ 1$	训练样本数	32000
波束宽度/(°)	5	验证样本数	4000
阵元间距	${\lambda \mathord{\left/ {\vphantom {\lambda 2}} \right. } 2}$	测试样本数	4000
来波方向范围/$(^\circ) $	$[ - 3 ,3]$	过拟合因子dropout	0.5
量化单元$\Delta \phi/(^\circ) $	$ 0.1 $	空域（场景）数	6

模型	输入信息维度	RMSE/(°)
AISL（未稀疏降维） AISL（稀疏降维）	420 160	0.1679 0.1506

[1]	杨宗方, 张天骐, 马焜然, 邹涵. 基于深层神经网络的信道编码类型盲识别[J]. 系统工程与电子技术, 2024, 46(5): 1820-1829.
[2]	霍超颖, 闫华, 冯雪健, 殷红成, 邢笑宇, 陆金文. HRRP稀疏自编码器深层特征与散射中心特征的关联性研究[J]. 系统工程与电子技术, 2021, 43(11): 3040-3053.