闪烁现象下旋翼目标微动参数估计方法

doi:10.12305/j.issn.1001-506X.2022.01.08

摘要/Abstract

摘要：

逆Radon变换以其精度高、抗噪性能好的优点常用于微动信号的参数估计, 但是当旋翼类目标微动信号存在闪烁现象时, 该方法失效。针对此问题, 提出一种闪烁现象下的微动参数估计方法。首先, 建立基于线性调频信号的单旋翼直升机雷达回波散射点模型, 分析闪烁现象下回波的微动特性。其次, 通过去噪卷积神经网络(denosing convolutional neural network, DnCNN)结构分别训练去噪网络和去闪烁网络, 消除旋翼目标回波时频图中存在的噪点、闪烁带和零频带, 得到余弦包络特征增强的微动信号时频图。最后, 针对传统逆Radon变换使用遍历法搜索微动参数, 存在运算量较大的问题, 因此采用黄金分割法对搜索过程进行改进, 提升参数估计速度, 最终完成对旋翼目标微动参数的估计。仿真结果验证了所提方法的可行性和有效性。

关键词: 旋翼目标, 微多普勒, 时频分析, 深度学习, 逆Radon变换, 黄金分割法, 参数估计

Abstract:

Inverse Radon transform is often used to estimate the parameters of micro-motion signals because of its high precision and good denoising performance. However, when the micro-motion signal of the rotor target has a flashing phenomenon, the method fails. In order to solve this problem, a method to estimate the micro-motion parameters under the flashing phenomenon is proposed. Firstly, the scattering point model of single-rotor helicopter radar echoes based on linear frequency modulation signals is established, and the micro-motion characteristics of the echoes under the flashing phenomenon are analyzed. Secondly, the denoising network and the deflashing network are trained respectively through the denosing convolutional neural network(DnCNN) structure to eliminate the noise, flashing band and zero band in the time-frequency diagram of rotor target echoes, and the time-frequency diagram of micro-motion signals enhanced by cosine envelope feature is obtained. Finally, the traditional inverse Radon transform uses the ergodic method to search for micro-motion parameters, which has a large amount of calculation. Therefore, the golden section method is adopted to improve the search process and the speed of parameter estimation, and finally complete the estimation of micro-motion parameters of the rotor target. Simulation results verify the feasibility and effectiveness of the proposed method.

Key words: rotor target, micro-Doppler, time-frequency analysis, deep learning, inverse Radon transform, golden section method, parameter estimation

中图分类号:

TN957.51

周毅恒, 杨军, 夏赛强, 吕明久. 闪烁现象下旋翼目标微动参数估计方法[J]. 系统工程与电子技术, 2022, 44(1): 54-63.

Yiheng ZHOU, Jun YANG, Saiqiang XIA, Mingjiu LYU. Estimation method of micro-motion parameters for rotor targets under flashing[J]. Systems Engineering and Electronics, 2022, 44(1): 54-63.

图/表 22

图1

图2

图3

图4

图5

图6

图7

图8

表1

表2

图9

图10

图11

表3

图12

图13

表4

表5

表6

表7

表8

表9

参考文献 27

1	CHEN V C. Analysis of radar micro-Doppler signature with time-frequency transform[C]//Proc. of the IEEE 10th Workshop on Statistical Signal and Array Processing, 2000: 463-466.
2	杜兰, 李林森, 李玮璐, 等. 基于时域回波相关性特征的飞机目标分类方法[J]. 雷达学报, 2015, 4 (6): 621- 629.
	DU L , LI L S , LI W L , et al. Aircraft target classification based on correlation features from time-domain echoes[J]. Journal of Radars, 2015, 4 (6): 621- 629.
3	姜悦, 范菊平, 郭乐田, 等. 基于时频图的飞机目标特征提取算法[J]. 现代雷达, 2016, 38 (4): 38- 41.
	JIANG Y , FAN J P , GUO L T , et al. Feature-extraction algorithm of airplane targets based on time-frequency diagram[J]. Modern Radar, 2016, 38 (4): 38- 41.
4	章鹏飞, 李刚, 霍超颖, 等. 基于双雷达微动特征融合的无人机分类识别[J]. 雷达学报, 2018, 7 (5): 557- 564.
	ZHANG P F , LI G , HUO C Y , et al. Classification of drones based on micro-Doppler radar signatures using dual radar sensors[J]. Journal of Radars, 2018, 7 (5): 557- 564.
5	STANKOVIC L , DAKOVIC M , THAYAPARAN T , et al. Inverse radon transform-based micro-Doppler analysis from a reduced set of observations[J]. IEEE Trans.on Aerospace and Electronic Systems, 2015, 51 (2): 1155- 1169. doi: 10.1109/TAES.2014.140098
6	ZHANG Q , YE T S , TAN H S , et al. Imaging of a moving target with rotating parts based on the Hough transform[J]. IEEE Trans.on Geoscience and Remote Sensing, 2008, 46 (1): 291- 299. doi: 10.1109/TGRS.2007.907105
7	张群, 胡健, 罗迎, 等. 微动目标雷达特征提取、成像与识别研究进展[J]. 雷达学报, 2018, 7 (5): 531- 547.
	ZHANG Q , HU J , LUO Y , et al. Research progresses in radar feature extraction, imaging, and recognition of target with micro-motion[J]. Journal of Radars, 2018, 7 (5): 531- 547.
8	夏赛强, 向龙, 朱名烁, 等. 一种变步长的微动目标参数高精度提取方法[J]. 雷达科学技术, 2019, 17 (5): 506- 512.
	XIA S Q , XIANG L , ZHU M S , et al. A variable-step high accuracy extraction algorithm for micro-motion target parameters[J]. Radar Science and Technology, 2019, 17 (5): 506- 512.
9	马娇, 董勇伟, 李原, 等. 多旋翼无人机微多普勒特性分析与特征提取[J]. 中国科学院大学学报, 2019, 36 (2): 235- 243.
	MA J , DONG Y W , LI Y , et al. Multi-rotor UAV's micro-Doppler characteristic analysis and feature extraction[J]. Journal of University of Chinese Academy of Sciences, 2019, 36 (2): 235- 243.
10	陈广锋, 张林让, 刘高高. 基于微多普勒分析的直升机旋翼参数估计[J]. 计算机工程, 2012, 38 (17): 249- 253.
	CHEN G F , ZHANG L R , LIU G G . Parameter estimation of helicopter blade based on micro-Doppler analysis[J]. Computer Engineering, 2012, 38 (17): 249- 253.
11	陈永彬, 李少东, 杨军, 等. 旋翼叶片回波建模与闪烁现象机理分析[J]. 物理学报, 2016, 65 (13): 287- 297.
	CHEN Y B , LI S D , YANG J , et al. Rotor blades echo modeling and mechanism analysis of flashes phenomena[J]. Acta Physica Sinica, 2016, 65 (13): 287- 297.
12	夏勇, 田西兰, 常沛, 等. 基于深度学习的复杂沙漠背景SAR目标检测[J]. 雷达科学与技术, 2019, 17 (3): 305- 309, 318. doi: 10.3969/j.issn.1672-2337.2019.03.011
	XIA Y , TIAN X L , CHANG P , et al. SAR target detection in complex desert background images based on deep learning[J]. Radar Science and Technology, 2019, 17 (3): 305- 309, 318. doi: 10.3969/j.issn.1672-2337.2019.03.011
13	WANG F , SHENG S C . Residual learning of deep convolutional neural network for seismic random noise attenuation[J]. IEEE Geoscience and Remote Sensing Letters, 2019, 16 (8): 1314- 1318. doi: 10.1109/LGRS.2019.2895702
14	秦剑. 基于生成对抗网络的信号重构[D]. 西安: 西安电子科技大学, 2018.
	QIN J. Signal reconstruction based on generation countermea-sure network[D]. Xi'an: Xidian University, 2018.
15	韩萍, 孙丹丹. 特征选择与深度学习相结合的极化SAR图像分类[J]. 信号处理, 2019, 35 (6): 972- 978.
	HAN P , SUN D D . Polarimetric SAR image classification based on feature selection and deep learning[J]. Signal Processing, 2019, 35 (6): 972- 978.
16	ZHANG K , ZUO W , CHEN Y , et al. Beyond a Gaussian denoiser: residual learning of deep CNN for image denoising[J]. IEEE Trans.on Image Processing, 2017, 26 (7): 3142- 3155. doi: 10.1109/TIP.2017.2662206
17	XIA X G . System identification using chirp signals and time-variant filters in the joint time-frequency domain[J]. IEEE Trans.on Signal Processing, 1997, 45 (8): 2072- 2084. doi: 10.1109/78.611210
18	夏赛强, 向虎, 陈文峰, 等. 基于CEMD的旋翼微动目标杂波抑制方法[J]. 航空学报, 2018, 39 (9): 184- 193.
	XIA S Q , XIANG H , CHEN W F , et al. Clutter suppression method for rotor micro-motion target based on CEMD[J]. Acta Aeronautica et Astronautica Sinica, 2018, 39 (9): 184- 193.
19	王才, 徐成发, 冯祺, 等. 基于瞬时频率率的平动刚体目标微动参数测量[J]. 北京理工大学学报, 2017, 37 (3): 318- 324.
	WANG C , XU C F , FENG Q , et al. Micro-motion parameter measurement of rigid targets with translation based on instantaneous frequency rate[J]. Transactions of Beijing Institute of Technology, 2017, 37 (3): 318- 324.
20	王雨千, 周峰, 王世强. 弹道目标微多普勒特性基于逆Radon变换的稀疏表示[J]. 火力与指挥控制, 2017, 42 (9): 55- 59. doi: 10.3969/j.issn.1002-0640.2017.09.012
	WANG Y Q , ZHOU F , WANG S Q . Sparse representation of micro-Doppler feature of ballistic target based on inverse Radon transform[J]. Fire Control & Command Control, 2017, 42 (9): 55- 59. doi: 10.3969/j.issn.1002-0640.2017.09.012
21	IOFFEAND S, SZEGEDY C. Batch normalization: accelerating deep network training by reducing internal covariate shift[C]//Proc. of the Internationa Conference on Machine Learning, 2015: 448-456.
22	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2016.
23	李昌利, 胡丽娜. SAR微动目标检测及其参数估计方法[J]. 雷达科学与技术, 2019, 17 (4): 365- 370. doi: 10.3969/j.issn.1672-2337.2019.04.003
	LI C L , HU L N . An algorithm for micro motion target detection and parameter estimation for synthetic aperture radar[J]. Radar Science and Technology, 2019, 17 (4): 365- 370. doi: 10.3969/j.issn.1672-2337.2019.04.003
24	赵彤璐, 廖桂生, 杨志伟. 基于短时迭代自适应-逆Radon变换的微多普勒提取方法[J]. 电子学报, 2016, 44 (3): 505- 513. doi: 10.3969/j.issn.0372-2112.2016.03.002
	ZHAO T L , LIAO G S , YANG Z W . Micro-Doppler extraction based on short-time iterative adaptive approach and inverse Radon transform[J]. Acta Electronica Sinica, 2016, 44 (3): 505- 513. doi: 10.3969/j.issn.0372-2112.2016.03.002
25	王童, 童创明, 李西敏, 等. 扩展性微动目标回波模拟与特征参数提取研究[J]. 物理学报, 2015, 64 (21): 168- 176.
	WANG T , TONG C M , LI X M , et al. Echo simulation and feature parameter extraction of extended fretting target[J]. Acta Physica Sinica, 2015, 64 (21): 168- 176.
26	李康乐, 刘永祥, 姜卫东, 等. 基于逆Radon变换的微动目标重构研究[J]. 雷达科学与技术, 2010, 8 (1): 74- 79, 86. doi: 10.3969/j.issn.1672-2337.2010.01.015
	LI K L , LIU Y X , JIANG W D , et al. Reconstruction of moving targets based on inverse Radon transform[J]. Radar Science and Technology, 2010, 8 (1): 74- 79, 86. doi: 10.3969/j.issn.1672-2337.2010.01.015
27	TAUSSKY O , KNUTH D E . The art of computer programming(Ⅱ): seminumerical algorithms[J]. The American Mathematical Monthly, 1970, 77 (8): 900. doi: 10.2307/2317055

参数	参数值
翼长l/m	5~12之间任意数
旋翼个数I	2~6个之间任意数
转动角频率ω/(rad/s)	6π~14π之间任意数
俯仰角β/(°)	20~60之间任意数
叶片的散射点个数N	70(生成集合S₄时为2)
方位角α/(°)	0

参数	参数值
窗长	100
频率点数	2 048
采样频率/Hz	4 000

类型	叶片数	叶片长度m	旋转角频率rad/s	叶片1的初相角/(°)	叶片2的初相角/(°)	叶片3的初相角/(°)	叶片4的初相角/(°)	俯仰角βrad
图 9(a)	3	8.42	19.24	120	240	360	-	0.41
图 10(a)	2	10.44	23.33	180	360	-	-	0.59
图 11(a)	4	7.63	29.37	90	180	270	360	0.45

类型	实际旋翼长度	叶片1长度估计值	叶片2长度估计值	叶片3长度估计值	叶片4长度估计值	叶片长度估计均值
图 9(c)	8.42	9.27	8.21	7.71	-	8.40
图 10(c)	10.44	11.16	10.27	-	-	10.72
图 11(c)	7.63	7.46	7.08	8.07	7.72	7.58

类型	实际转动角频率(rad/s)	角频率估计值(rad/s)	误差%
图 9(c)	19.24	19.38	0.7
图 10(c)	23.33	23.28	0.2
图 11(c)	29.37	29.58	0.7