基于时延估计波束形成预处理的实时语音盲信号分离研究  

On Real-time Speech Blind Signal Separation Based on Time Delay Estimation and Beamforming

作　　者：吴景田[1] 何培宇[1] 邓方[1] 徐自励[1] 潘帆[1] 

机构地区：[1]四川大学电子信息学院,成都610064

出　　处：《四川大学学报（自然科学版）》2006年第1期99-103,共5页Journal of Sichuan University(Natural Science Edition)

基　　金：国家自然科学基金(60472096)

摘　　要：针对现有盲信号分离算法,在真实环境下性能会降低及计算量大而难以实时处理的局限性,提出一种将基于时延估计的波束形成与一种简单的去相关盲分离算法相结合的分离方法.在波束形成中,首先用互功率谱相位法对信号源到达传感器阵列的相对时延进行估计,然后进行延时对消,在此基础上再进行盲信号分离,并将该去相关盲分离算法推广到频域进一步降低计算复杂度.仿真实验表明,该方法简单可行并使得分离效果显著改善.Authors have addressed the problem of speech blind signal separation （BSS） in real-time by using a new approach that combines the beamforming and a decorrelation algorithm for BSS. In the stage of beamforming, time delay estimation and time delay compensation are performed through the Cross-power Spectrum Phase method. Furthermore, a conventional decorrelation algorithm is used to separate the signals. We validate the simplicity of the method and reveal that the signal separation performance of the proposed method is superior to that of the conventional decorrelation-based BSS method through the experiments of simulation.

关 键 词：盲信号分离 波束形成 互功率谱相位 时延估计 

分 类 号：TN911.7[电子电信—通信与信息系统]