多子阵高分辨实时波达估计算法研究  

Real-time high-resolution DOA estimation based on sub array partition

作　　者：田彪[1,2] 黄海宁[1] 叶青华[1] 李宇[1] 

机构地区：[1]中国科学院声学研究所,北京100190 [2]中国科学院研究生院,北京100190

出　　处：《仪器仪表学报》2010年第3期513-518,共6页Chinese Journal of Scientific Instrument

基　　金：中国科学院创新工程国防科研项目;知识创新工程青年人才领域前沿项目;国家自然科学基金项目(60602054)资助

摘　　要：常用的高分辨波达估计算法一般都面临计算量巨大的高维协方差矩阵求逆或需要已知目标数目的先验信息等难题,难以在实际中应用。本文从多子阵处理策略出发,结合最小无失真响应算法(MVDR)理论,提出了一种自适应高分辨实时波达估计算法,并给出了详细的理论推导及相应的性能分析。研究表明,本文算法波达估计误差优于频域常规波束形成算法,同时,相较于常用算法实现高分辨波达估计所需的10 dB以上的阵元信噪比门限,本文算法所需的信噪比门限为3.5 dB,更适用于低信噪比条件下目标的探测,并且对舷侧阵阵元信噪比分布不一致性有较强的宽容性。更重要的是,由于协方差矩阵维数的大大降低,使得本文算法的处理速度大大提升,从而使得高分辨的实时波达估计成为可能。仿真及海试数据处理验证了本文算法的可行性。Many high resolution direction-of-arrival（DOA） estimation approaches have been studied to resolve close targets,which include the Minimum Variance Distortionless Response（MVDR）,MUSIC,ESPRIT,and so forth. However there are some difficult problems need to be dealt with in real applications like huge dimension matrix inversion, source number estimation,broadband source detection and so on.A real-time high-resolution DOA estimation algorithm based on sub array partition is developed in details,which is based on sub array partition and MVDR algorithm. The proposed algorithm can be applied to broadband source detection and has high resolution ability,what is more important is that it can be implemented in real-time,thus it is suitable for realtime applications.The validity of the proposed algorithm is verified by simulation results and real trial data.The result shows that its DOA estimation error outperforms those of the frequently used conventional beamforming algorithms;and the proposed algorithm has lower SNR threshold and is more robust to inconsistent SNR on sensors.

关 键 词：波达估计 子阵处理 MVDR 信噪比不一致 宽容性 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]