摘要
采用基于Jacobi算法的矩阵联合对角化方法对阵列接收数据的空时相关矩阵进行联合对角化,能够获取关于目标方位的尽可能多的信息,即空时相关矩阵"平均"特征值和特征向量,将这些平均值应用于Capon提出的最小方差无畸变响应(MVDR)空间谱估计算法,得到修正MVDR算法。由于修正MVDR算法充分利用了目标方位信息,且不需要信号源数目的先验信息,所以,其性能优于MVDR算法和MUSIC算法。仿真试验表明,在信号源相关程度比较高的情况下,修正MVSDR算法不需要信源数目的先验信息,仍能分辨出信号方位,其性能明显优于MUSIC算法,特别是当信源数目欠估计时,而其他基于子空间的算法得不出任何有意义的结果。修正MVDR算法空时相关矩阵联合对角化能提高空间谱的分辨能力,在一定程度上增加算法的稳健性。
The spatio-temporal correlation matrix of array receiving data is joint block-diagonalized using the Jacobi algorithm to get the average eigenvalue and eigenvector. The average eigenvalue and eigenvector is used in the minimum variance distortionless response （MVDR） algorithm proposed by Capon to get the modified MVDR algorithm. Simulation example shows that the modified MVDR algorithm is better than the conventional MVDR algorithm. The performance of the modified MVDR algorithm is much better than MUSIC algorithm when the signal sources if partialy correlated and especially in the case of under-estimation of the number of signal sources. The prior information of the number of signal sources is not needed when using the modified MVDR algorithm.
出处
《控制工程》
CSCD
北大核心
2009年第5期602-605,共4页
Control Engineering of China
基金
国家863计划基金资助项目(2007ADA299)
关键词
时空相关矩阵
联合对角化
空间谱分解
相关信号源
最小方差无畸变响应
spatio-temporal correlation matrix
joint block-diagonalization
spatial spectral resolution
correlated sources, minimum variance distnrtionless response（MVDR）
