一种改进的声测定位时延估计算法  

An improved time-delay estimation algorithm for acoustic detection and location

作　　者：孙韶杰[1] 孙绍俊[2] 李国辉[1] 李红梅[2] 

机构地区：[1]国防科学技术大学信息系统与管理学院,湖南长沙410073 [2]山东省图书馆网络部,山东济南250100

出　　处：《计算机应用》2006年第11期2696-2697,2723,共3页journal of Computer Applications

摘　　要：研究了时延估计算法在被动声测定位中的应用,提出了一种改进的基于最大似然(ML)权函数的广义互相关时延估计算法。改进的算法采用加窗法和最小均方差(LMS)滤波法,弥补了原算法计算量大及无法消除回响干扰的不足。仿真结果表明,改进的算法计算复杂度明显降低,能够有效地消除回响干扰,具有较高的时延估计精度和鲁棒性。The application of time delay estimation in passive acoustic detection and location was studied. An improved maximum likelihood estimator as the generalized correlation method for time-delay estimation, which adopted the methods of windowing and least mean square （LMS） filter, was proposed. The simulation results show that the proposed algorithm can reduce the complexity of computation obviously, eliminate the influence of reecho efficiently, and has better estimation accuracy and robustness.

关 键 词：时延估计 广义互相关法 最大似然权函数 声测定向 最小均方差滤波 

分 类 号：TN911.72[电子电信—通信与信息系统]