基于双重Toeplitz矩阵填充的阵元受损MIMO雷达DOA估计  

DOA Estimation Based on Doubly Toeplitz Matrix Completion in MIMO Radar Under Antenna Failure

作　　者：陈金立[1] 张盈盈 李家强[1] CHEN Jin-li;ZHANG Ying-ying;LI Jia-qiang(School of Electronic and Information Engineering,Nanjing University of Information Science and Technology,Nanjing 210044,China)

机构地区：[1]南京信息工程大学电子与信息工程学院,江苏南京210044

出　　处：《中国电子科学研究院学报》2023年第4期301-308,共8页Journal of China Academy of Electronics and Information Technology

基　　金：国家自然科学基金资助项目(62071238);江苏省自然科学基金资助项目(BK20191399)。

摘　　要：在阵元受损下多输入多输出(Multiple Input Multiple Out,MIMO)雷达的虚拟阵列中存在虚拟阵元部分缺失,使得传统波达方向(Direction of Arriual,DOA)估计方法无法有效估计目标DOA。矩阵填充可以利用随机分布的少量采样元素恢复出完整矩阵,但阵元受损使得虚拟阵列协方差矩阵会出现结构性缺失数据,导致传统矩阵填充方法失效。为此,文中利用阵列协方差矩阵的双重Hermite Toeplitz结构和低秩特性,构建一种改进的低秩矩阵填充模型,并分别使用凸优化工具箱(CVX)和交替方向乘子法求解该模型获得完整的协方差矩阵,最后利用RD-ESPRIT算法获得目标的DOA估计值。仿真结果表明,文中的方法能对阵元受损带来的结构性缺失元素进行有效重构,从而提高了目标DOA估计性能。Some virtual array elements are missing in the virtual array of MIMO radar under antenna failures.This leads to the performance degradation for traditional DOA estimation methods.Matrix completion can recover the complete matrix by using a small number of randomly distributed sampling entries.However,the covariance matrix is structurally-incomplete due to the antenna failure,which causes the traditional matrix completion method fail to recover these missing data.To remedy this,the improved low-rank matrix completion model by using the double Hermite Toeplitz structure and low-rank property of the MIMO radar covariance matrix is formulated,and then the convex optimization toolbox(CVX) and the alternating direction method of multipliers are employed to solve this model,respectively.Finally,the DOAs of the targets are estimated by RD-ESPRIT algorithm.Simulation results show that the proposed method can effectively reconstruct the missing structural entries caused by the damaged antennas using the prior structure information of the MIMO radar covariance matrix,and improve the target DOA estimation performance.

关 键 词：阵元受损 MIMO雷达 矩阵填充 双重Toeplitz矩阵 RD-ESPRIT算法 

分 类 号：TN911.23[电子电信—通信与信息系统]