一种针对单快拍DOA估计的子空间搜索近似消息传递算法  

A Subspace Searching Approximation Message Passing Algorithm for Single Snapshot DOA Estimation

作　　者：曾令豪 刘静 韩崇昭[1] ZENG Ling-Hao1 ,LIU Jing1 ,HAN Chong-Zhao1(1. Ministry of Education Key Laboratory for Intelligent Net- works and Network Security, Xi＇an Jiaotong University, Xi＇an 71004)

机构地区：[1]西安交通大学智能网络与网络安全教育部重点实验室,西安710049

出　　处：《自动化学报》2018年第3期443-452,共10页Acta Automatica Sinica

基　　金：国家重点基础研究发展计划(973计划)(2013CB329405);国家自然科学基金(61221063;61573271;61573276;61370037)资助~~

摘　　要：压缩感知(Compressed sensing,CS)技术应用于单快拍波达方向(Direction of arrival,DOA)估计中可以实现相关信号的超分辨估计,但会遇到感知矩阵高相干性以及对噪声敏感的问题.本文提出一种基于近似消息传递的子空间搜索算法以解决上述问题.该算法首先通过近似消息传递算法得到一个粗解,随后利用该粗解划分子空间,最后在子空间中寻找精确解.仿真结果验证了所提算法的有效性.文章最后通过理论分析了该算法性能并讨论了算法在信号数未知时的扩展应用.When compressed sensing （CS） is applied to single snapshot direction of arrival （DOA） estimation, super- resolution reconstruction of the correlated signals becomes possible. However, problems such as high coherent sensing matrix and noise sensitivity come into being at the same time. In order to solve these problems, an approximate message passing based subspace search algorithm is proposed. Firstly, a rough solution is obtained by approximate message passing algorithm. Secondly, subspaces are divided according to the rough solution. Thirdly, an exact solution is found by searching the subspaces. Simulation results show the effectiveness of the proposed approach. Finally, the performance of the proposed algorithm is analzed, and the application in which the number of signals is unknown is discussed as well the performance.

关 键 词：DOA估计 压缩感知 近似消息传递算法 子空间搜索 

分 类 号：TN911.7[电子电信—通信与信息系统]