一种累积量线性预测方位估计的快速算法  

A Fast Algorithm of DOA Estimation Based on Fourth-Order Cumulant Linear Prediction

作　　者：冯西安[1] 雷济声[1] 黄建国[1] 

机构地区：[1]西北工业大学航海工程学院,陕西西安710072

出　　处：《西北工业大学学报》2001年第4期579-582,共4页Journal of Northwestern Polytechnical University

摘　　要：提出了一种基于四阶累积量的线性预测方位估计快速算法。通过对累积量线性预测基本方程进行变换 ,得到了基本方程系数矩阵的 Toeplitz结构 ,并用 Levinson- Durbin算法求解方程 ,减少了运算量。在线性预测方程中 ,使用了四阶累积量 ,有效地拟制了高斯噪声。计算机仿真结果表明了这种方法的有效性。High order cumulant linear prediction equations can be solved by SVD-TLS (singular value decomposition total least square) algorithm for DOA (direction-of-arrival) estimation; the computation required is O(M 3) , where M represents the number of array units. We propose a fast algorithm of DOA estimation based on fourth-order cumulant linear prediction. We obtained Toeplitz structure of the coefficient matrix of fourth-order cumulant linear prediction equations through transforming the equations and suggest solving the Toeplitz structure by Levinson-Durbin algorithm to reduce computation required to O((M-1) 2) . We applied fourth-order cumulant to reduce Gaussian noise. Computer simulation results, as given in Figs.3 through 6, show preliminarily that the proposed fast algorithm is effective.

关 键 词：四阶累积量 线性预测-方位估计 快速算法 高阶统计量 高斯噪声 信号处理 

分 类 号：TN911.7[电子电信—通信与信息系统]