电子学报 ›› 2021, Vol. 49 ›› Issue (12): 2315-2322.DOI: 10.12263/DZXB.20210347

• 学术论文 • 上一篇    下一篇

冲击噪声下基于张量分解和K‑means聚类的MIMO雷达阵列诊断

陈金立1,2, 王亚鹏2, 李家强1,2(), 龙伟军3()   

  1. 1.南京信息工程大学气象灾害预报预警与评估协同创新中心,江苏 南京 210044
    2.南京信息工程大学电子与信息工程学院,江苏 南京 210044
    3.南京电子技术研究所,江苏 南京 210039
  • 收稿日期:2021-03-12 修回日期:2021-04-22 出版日期:2021-12-25 发布日期:2021-12-25
  • 作者简介:陈金立 男,1982年8月出生于浙江宁波市. 现为南京信息工程大学副教授、硕士生导师. 研究方向为MIMO雷达信号处理,阵列信号处理.
    王亚鹏 男,1995年5月出生于江苏省盐城市,硕士研究生,主要研究方向为雷达阵列故障阵元诊断.E‑mail: wangyapeng9505@163.com
    李家强 男,1976年5月出生于安徽省滁州市.现为南京信息工程大学副教授,从事信号检测与估计和雷达系统设计的研究工作. E‑mail: lijiaqiang@sina.com
    龙伟军 男,1979年1月出生于重庆市合川区. 现为南京电子技术研究所研究员,从事雷达系统理论研究与工程应用的研究工作. E‑mail: chinacohit2020@163.com
  • 基金资助:
    国家自然科学基金(62071238);江苏省自然科学基金(BK20191399)

Tensor Decomposition and K‑means Clustering Based Array Diagnosis for MIMO Radar in Impulsive Noise Environment

CHEN Jin-li1,2, WANG YA-peng2, LI Jia-qiang1,2(), LONG Wei-jun3()   

  1. 1.Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters,Nanjing University of Information Science and Technology,Nanjing,Jiangsu 210044,China
    2.School of Electronic and Information Engineering,Nanjing University of Information Science and Technology,Nanjing,Jiangsu 210044,China
    3.Nanjing Research Institute of Electronics Technology,Nanjing,Jiangsu 210039,China
  • Received:2021-03-12 Revised:2021-04-22 Online:2021-12-25 Published:2021-12-25

摘要:

针对冲击噪声下多输入多输出(Multiple?Input Multiple?Output, MIMO)雷达阵列诊断失效问题,对基于二阶矩的传统匹配滤波器进行改进以适应非高斯噪声,并提出一种基于张量分解和K?means聚类的阵列诊断方法. 该方法利用MIMO雷达各接收阵元回波信号的高斯核函数值来自适应地调整匹配滤波器的系数,以有效形成虚拟阵列. 为挖掘正常和故障阵元的匹配滤波输出数据的多维特征,将虚拟阵列协方差矩阵构建成三阶平行因子(PARAllel FACtor, PARAFAC)张量,并通过COMFAC(COMplex parallel FACtor analysis)算法分解获得收发阵列流形矩阵,使用欧式距离度量其相似性,确定两个簇类数据的聚类中心并划分出异常簇类,以完成故障阵元位置的诊断. 仿真结果验证了所提算法的有效性.

关键词: MIMO雷达, 阵列诊断, 冲击噪声, 匹配滤波, 张量分解, K?means聚类

Abstract:

The traditional array diagnosis methods for multiple-input multiple-output(MIMO) radar may fail in the presence of impulse noise. The traditional matched filter based on second-order statistics is modified to obtain a reliable performance in the non-Gaussian noise, and then the array diagnosis method based on tensor decomposition and K-means clustering is proposed. The coefficients of the matched filters are adjusted with the Gaussian kernel function values of the echo signal observed at each receive element, which makes the MIMO radar form a virtual array successfully in the presence of impulsive noise. To further utilize the inherent multidimensional structure of the matched filter output data of the damaged and normal antennas, a third-order parallel factor(PARAFAC) model of the virtual array covariance matrix is formulated. By exploiting the complex parallel factor analysis(COMFAC) algorithm on the third-order covariance tensor, the manifold matrices of the transmit and receive arrays are obtained. The similarity of manifold matrix data is measured using Euclidean distance, and the clustering centers of the two clusters corresponding to the normal and fault elements are determined. The abnormal cluster data is selected to diagnose the location of the fault elements in MIMO radar array. Numerical simulation results confirm the effectiveness of the proposed algorithm.

Key words: MIMO radar, array diagnosis, impulsive noise, matched filter, tensor decomposition, K-means clustering

中图分类号: