阵列高分辨方位估计方法统计性能分析

STATISTICAL PERFORMANCE COMPARISON OF HIGH RESOLUTION DOA ESTIMATION METHODS

在简要回顾近年来阵列高分辨定向技术发展的基础上，从算法机理出发，对目前出现的各种高分辨定向算法进行合理分类。重点对一批典型算法的多目标定向性能进行统计分析，并利用统一的仿真模型对它们的估计精度和分辨能力进行比较，结论为：ＭＵＳＩＣ，Ｍｉｎｉ－Ｎｏｒｍ两种方法的实用性较强，而ＷＳＦ，ＩＭＰ，ＴＬＳ－ＥＳＰＲＩＴ等方法具有较高的理论研究价值，有待进一步开拓其实用性。

A wide comparison is presented of high resolution Direction of Arrival (DOA) methods. The development of DOA estimation approaches is reviewed briefly. In order to study their performance clearly, more than 30 methods were divided into 5 groups, which are conventional methods, parametric modeling methods, subspace based methods, deconvolution methods, and other methods. Ten typical methods of them were studied in resolution and precision respectively. Based on a unitized simulation model, they were compared through statistical analysis. Furthermore, several criteria for application were proposed to judge their practicality. At last, MUSIC, Mini Norm are considered to be worth the most promising methods for utility while WSF, etc. to be worth theoretical studying.