基于最小熵值的麦克风阵列声源定位算法  

Microphone Array Acoustic Source Localization Algorithm Based on Minimum Entropy

作　　者：刘颖[1,2] 刘建平[2] 夏靖波[1] 

机构地区：[1]空军工程大学电讯工程学院,西安710077 [2]武警工程学院通信工程系,西安710086

出　　处：《计算机工程》2012年第7期145-147,共3页Computer Engineering

摘　　要：针对传统麦克风阵列声源定位算法抗噪声及混响能力不强的问题,提出一种基于最小熵值和随机域压缩的麦克风阵列声源定位算法。利用最小熵值方法对麦克风阵列进行时延估计,并与随机域压缩方法相结合,对声源进行空间搜索。仿真实验结果表明,该算法在定位精度、抗噪声及抗混响能力方面均优于广义互相关-相位变换算法。Accurate localization of acoustic sources in high noise and reverberation environment is a problem for traditional source localization algorithm.This paper proposes a novel acoustic source localization algorithm for microphone array——Minimum Entropy-stochastic Region Contraction（ME-SRC）.The algorithm shows that acoustic source with Laplace distribution can be developed to estimate time delays between microphones on a basis of ME and SRC is used to localize the acoustic source in search space.Results show that the proposed ME-SRC algorithm is much more robust than GCC-PHAT in noise and reverberation environment.

关 键 词：声源定位 麦克风阵列 最小熵值 随机域压缩 拉普拉斯分布 时延估计 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]