声矢量传感器阵中基于Kalman滤波和OPASTd的DOA跟踪算法  

DOA Tracking Algorithm for Acoustic Vector-Sensor Array via Kalman Filter and OPASTd

作　　者：陈未央[1,2] 张小飞[1,2] 张立岑 

机构地区：[1]南京航空航天大学江苏省物联网与控制技术重点实验室,南京210016 [2]南京航空航天大学电子信息工程学院,南京210016

出　　处：《南京航空航天大学学报》2015年第3期377-383,共7页Journal of Nanjing University of Aeronautics & Astronautics

基　　金：国家自然科学基金(61371169)资助项目;江苏省博士后科研资助计划(1201039C)资助项目;中国博士后基金(2012M521099)资助项目;江苏高校优势学科建设工程资助项目

摘　　要：研究了声矢量传感器阵动目标角度跟踪问题,并提出了声矢量传感器阵中一种基于Kalman滤波和正交压缩近似投影子空间跟踪(Orthonormal projection approximation and subspace tracking of deflation,OPASTd)的波达方向(Direction of arrival,DOA)跟踪算法。该算法通过OPASTd算法来进行DOA的跟踪,从而克服了PASTd算法由于在某些情况下振荡但不收敛进而压缩数据、在迭代更新中由特征向量的不准确性产生误差累积等原因引起破坏信号子空间正交性的缺陷。Kalman滤波和OPASTd相结合算法可在估计角度的同时进行数据关联,与传统的PASTd算法相比,角度跟踪性能更好。该算法的优越性均可在文中得到验证。A direction of arrival（DOA）tracking algorithm via Kalman filter and orthonormal projection approximation and subspace tracking of deflation（OPASTd）for acoustic vector-sensor array is proposed based on the investigation of target tracking.The proposed algorithm uses OPASTd algorithm to track DOA,thus overcoming the destroyed orthogonality of signal subspace of PASTd algorithm.The reasons for this defelt are：Firstly,the vibrated and non-convergent PASTd algorithm leads to data compression in some cases;secondly,the PASTd algorithm produces error accumulation by the inaccuracy of eigenvectors during iterative update.The proposed algorithm can estimate angle and contact data,which has a better DOA tracking performance than traditional PASTd algorithm.Simulation results verify the usefulness of the proposed algorithm.

关 键 词：声矢量传感器阵 波达方向估计 KALMAN滤波 OPASTd 

分 类 号：TN911.7[电子电信—通信与信息系统]