摘要
扫描雷达由于天线孔径受到平台尺寸限制,存在角分辨率低下的缺陷。角超分辨技术能够在不改变雷达硬件系统的前提下,通过信号处理方法,获得超越雷达实孔径波束宽度的角度分辨能力。近些年研究表明,基于最小绝对收缩和选择算子(LASSO)的稀疏角超分辨方法能够获得比传统方法更高的分辨能力。然而,该方法基于高斯白噪声假设,在残余项中主要基于l2范数进行约束。当回波数据中存在重尾分布的噪声时,该方法中的l2范数约束无法有效抑制重尾分布噪声,导致目标分辨能力下降,甚至产生虚假目标。针对该问题,本文提出一种扫描雷达抗重尾噪声角超分辨成像方法。首先,本文引入一种最小绝对偏差(LAD)-LASSO约束准则以抑制重尾分布噪声;更进一步的,针对模型中存在的正则化参数自适应选取难题,本文基于稀疏自相关迭代准则导出正则化参数的最优表达。最后,针对LAD-LASSO非平滑代价函数最优化求解难题,本文提出一种基于迭代重加权最小二乘(IRLS)算法的求解方法,通过将l1范数替换为重加权的l2范数,并在每次迭代中计算权值实现最优求解。仿真结果表明,相比传统稀疏角超分辨方法,本文提出的方法在不同信噪比下均能够有效抑制重尾分布噪声,并有效提升扫描雷达角分辨率。同时,本文采用岸基X波段雷达实测数据验证了本文提出方法的有效性。
The antenna aperture of scanning radar systems is limited by the platform size.As a result,scanning radar sys⁃tems suffer from the disadvantage of low angular resolution.Angular super-resolution technology can obtain an angular resolution beyond the beam width determined by real aperture using signal processing method without changing the radar hardware system.Recent studies have shown that sparse angular super-resolution method based on minimum absolute shrinkage and selection operator(LASSO)can obtain higher resolution than traditional super-resolution methods.How⁃ever,this method is based on the white Gaussian noise hypothesis and imposes the l2-norm on the residual items.When the heavy-tailed noise exists in the echo data,the l2-norm constraint in the LASSO method cannot effectively suppress the heavy-tailed noise,resulting in degraded angular resolution and false targets.To solve this problem,this paper presents an angular super-resolution imaging method to suppress the heavy-tailed noise.First,a minimum absolute deviation(LAD)-LASSO constraint criterion is introduced to suppress the heavy-tailed distribution noise.Furthermore,the optimal regularization parameter is selected based on the covariance fitting criterion to solve the problem of adaptive selection of regularization parameters in the model.Finally,in order to solve the non-smooth cost function,an iterative weighted least squares(IRLS)algorithm is presented.The l1-norm is replaced by the iterative weighted l2-norm,and the weights are cal⁃culated in each iteration to achieve the optimal solution.The results of simulation show that the proposed method can effec⁃tively suppress heavy-tailed noise and improve the scanning radar angular resolution under varying signal-to-the-noise ratio.Meanwhile,real data set acquired by a shore-based X-band radar demonstrate the effectiveness of the proposed method.
作者
张寅
张永超
晏家楠
刘帅迪
杨建宇
ZHANG Yin;ZHANG Yongchao;YAN Jianan;LIU Shuaidi;YANG Jianyu(School of Information and Communication Engineering,University of Electronic Science and Technology of China,Chengdu,Sichuan 611731,China;Yangtze Delta Region Institute(Quzhou),University of Electronic Science and Technology of China,Quzhou,Zhejiang 324000,China)
出处
《信号处理》
CSCD
北大核心
2023年第9期1562-1572,共11页
Journal of Signal Processing
基金
四川省自然科学基金项目(2023NSFSC1970,2022NSFSC0950)
衢州市财政资助科研项目(2022D0011,2022D036)。
关键词
扫描雷达
角超分辨
重尾噪声
最小绝对偏差-最小绝对收缩和选择算子
迭代重加权最小二乘
scanning radar
angular super-resolution
heavy-tailed noise
minimum absolute deviation-minimum abso⁃lute shrinkage and selection operator
iterative weighted least squares
