摘要
阵列信号处理中,MUSIC、ESPRIT等高分辨率DoA估计算法都要通过特征值分解来获得波达方向估计,然而矩阵特征值分解的计算量较大,不利于实时处理。提出使用PCA(principal component analysis)高效迭代算法,来逼近信号子空间矢量。该算法的计算过程相对简单,并具有自组织特性,适合于神经网络实现。仿真结果表明,所提算法的DoA估计性能与MUSIC算法相当。
Many high resolution algorithms based on subspace technology like MUSIC and ESPRIT estimate the direction of arrival of plane wave impinging on an array of sensors via eigendecomposition in array signal processing. However high computational burden for eigendecomposition makes them unsuitable for real time processing. An iterative algorithm for signal subspace estimation based on principal component analysis is proposed. The algorithm contains only relatively simple operations and has self organizing properties. Simulations show that the proposed algorithm has an analogy performance with MUSIC algorithm.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2005年第8期1376-1378,共3页
Systems Engineering and Electronics
关键词
波达方向
主成分分析
多重信号分类
direction of arrival
principal component analysis
multiple signal classification