快速四阶累积量旋转不变子空间算法  

Estimation of Signal Parameters via Rotational Invariance Techniques Algorithm with Speedy Fourth-Order Cumulants

作　　者：唐建红[1] 司锡才[1] 彭巧乐 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,哈尔滨150001

出　　处：《西安交通大学学报》2009年第6期88-92,共5页Journal of Xi'an Jiaotong University

基　　金：国防基础科研基金资助项目(A2420061104-06);哈尔滨科技人才研究专项基金资助项目(2007RFQXG030).

摘　　要：针对空间谱估计算法同时分辨多目标需要较多阵元、测角精度低且运算量大的问题,提出一种快速四阶累积量旋转不变子空间(MFOC-E)算法.首先根据四阶累积量矩阵构成的规律,取出所有不重复的元素构成新矩阵,去除了原矩阵的数据冗余并降低了原矩阵维数,然后通过分析新虚拟阵列的阵元位置特性,找到与原真实阵列结构完全相同的新子阵,同时设定了适用于所有线阵的子阵划分方法,最后利用旋转不变子空间算法确定目标的来向.与传统的同类算法相比,MFOC-E算法具有不损失天线孔径、测角精度高且运算量小的特点.仿真实验表明,MFOC-E算法在阵元数极少时,仍可以快速准确地实现多目标的到达角估计,三元阵列时能节省30%的运算时间.Using spatial spectrum estimation methods to resolve multiple targets at the same time needs more antennas, and has low accuracy of angle measurement with high calculation cost. Focusing on these problems, an estimation of signal parameters via rotational invariance techniques （ESPRIT） algorithm with speedy fourth-order cumulants （MFOC-E） is proposed. All non-repeated elements of the fourth-order cumulants matrix are selected to compose a new matrix according to the composing law of the fourth-order cumulants matrix so that the data redundancy of the primary matrix is removed and the dimension is reduced. Then a new sub-array which is the same as the primary real array is found through analyzing the position characteristic of the new virtual antenna, and a dividing sub-arrays method that adapts to all linear arrays is designed. Finally, the ESPRIT algorithm is used to estimate the directions of arrival （DOA）. It is found by comparing with traditional similar algorithm that the MFOC-E algorithm reteins antenna aperture, and ensures high accuracy of angle measurement with lower computational cost. Simulation results show that the MFOC-E algorithm estimates DOA of more targets rapidly and exactly with less antennas, and saves 30% calculation time with only three antennas.

关 键 词：空间谱估计 四阶累积量 旋转不变子空间 虚拟阵列 

分 类 号：TN911.7[电子电信—通信与信息系统]