一种新的基于EEF准则的空间声源个数估计算法

发布时间：2018-01-11 13:12

  本文关键词：一种新的基于EEF准则的空间声源个数估计算法　出处：《电子学报》2016年03期 　论文类型：期刊论文

  更多相关文章： 空间声源个数估计 球麦克风阵列 指数嵌入族准则 最小均方差

【摘要】：针对指数嵌入族(Exponentially Embedded Families,EEF)准则在快拍数小于阵元数情况下无法估计声源个数的问题,本文提出一种新的空间声源个数估计算法.首先通过球麦克风阵列采集空间声场高阶信息,建立球阵列信号模型,将声源个数估计扩展到三维空间.继而将观测信号空间分解为信号子空间和噪声子空间,利用最小均方差(Minimum Mean-Squared Error,MMSE)方法估计观测信号空间及噪声子空间的协方差矩阵,确保矩阵估计的一致性和准确性.在此基础上改进似然比函数,同时引入新的自由度计算,使得算法在快拍数小于阵元数的情况下能有效估计声源个数.仿真结果表明,在进行空间声源个数估计时,相对于EEF准则,新的算法不仅适用于快拍数小于阵元数情况,同时提高了估计准确率.
[Abstract]:This paper aims at the problem that the number of sound sources can not be estimated under the condition that the fast beat number is less than the number of array elements, and the exponential embedding family Embedded families EEF criterion can not estimate the number of sound sources. In this paper, a new algorithm for estimating the number of spatial sound sources is proposed. Firstly, the high-order information of spatial sound field is collected by the spherical microphone array, and the model of the spherical array signal is established. The estimation of the number of sound sources is extended to three dimensional space, and then the observation signal space is decomposed into signal subspace and noise subspace. The covariance matrix of observation signal space and noise subspace is estimated by minimum mean square error Mean-Squared (MMSE) method. To ensure the consistency and accuracy of matrix estimation, the likelihood ratio function is improved and a new degree of freedom calculation is introduced. The algorithm can effectively estimate the number of sound sources when the number of beats is less than the number of array elements. The simulation results show that the estimation of the number of sound sources in space is relative to the EEF criterion. The new algorithm is not only suitable for the cases where the number of beats is less than the number of elements, but also improves the accuracy of estimation.
【作者单位】： 上海大学通信与信息工程学院;
【基金】：国家自然科学基金(No.61001160) 上海市教委创新基金(No.12YZ023) 上海市自然科学基金(No.14ZR141500)
【分类号】：TN911.7
【正文快照】： 1引言信源个数估计是阵列信号处理的首要环节,在雷达、声纳、勘探、通信和生物医学等多个领域都有广泛的应用.大多数信源个数估计算法都是以线阵和平面阵[1,2]为基础建立信号模型进行个数估计.近年来,球阵逐步受到人们的重视,由于具有三维空间的旋转对称结构和较高的空间分辨，

本文编号：1409653