针对冲击噪声下因接收信号二阶及以上矩不存在而产生性能恶化的问题, 提出一种基于QR分解和鲁棒性主成分分析法 (QR-RPCA)的双基地多输入多输出 (MIMO)雷达参数估计方法。针对RPCA算法适用于实数矩阵处理的情况, 先将复数信号转化为实数; 然后根据冲击噪声的稀疏特点与目标信号矩阵的低秩特点, 利用 QR-RPCA算法将低秩信号矩阵从受冲击噪声污染的接收信号中提取出来, 并直接得到信号子空间, 该算法避免了传统 RPCA算法中的大规模奇异值分解, 时间复杂度有所降低; 最后根据信号子空间并利用旋转不变信号参数估计技术 (ESPRIT)对目标方位进行估计。理论与仿真表明, 本文算法相较于其他消除冲击噪声的算法, 对于低特征指数的冲击噪声具有更好的估计性能。

In view of the performance degradation in impulsive noise environment due to the non-existence of the second-order moment, an estimation algorithm based on orthogonal trigonometric(QR)-Robust Principal Component Analysis(QR-RPCA) in bi-static ultiple-Input-Multiple-Output(MIMO) radar is proposed. At first, the complex received signal is converted into real-value for the RPCA algorithm. After that, according to the sparsity of impulsive noise and the low rank of signal matrix, the QR-RPCA algorithm is utilized to extract the low-rank signal matrix from the received signal polluted by impulsive noise and acquire the signal subspace directly. The QR-RPCA algorithm averts solving large-scale Singular Value Decomposition(SVD) and reduces the complexity relative to other traditional algorithms. Finally, Estimating Signal Parameters via Rotational Invariance Techniques(ESPRIT) are adopted to obtain the target locations. The theory and the simulation show that the proposed method has better estimation performance than the traditional ones in impulsive noise elimination while the characteristic index of impulsive noise is low.