使用多重訊號分類演算法之多聲源方位追蹤與分離

Abstract

本論文提出一套利用麥克風陣列偵測聲源方位並進行追蹤的方法，本方法利 用多重訊號分類(multiple signal classification)演算法估測聲源方位， 並對聲源頻譜進行估測。結合聲源頻譜與空間分佈資訊，利用機率決策方法 對未知數量聲源方位進行追蹤。由於本方法結合了聲源的空間與頻率特性， 可在低訊噪比下對多聲源進行追蹤並保有強健的偵測率，且可排除錯誤偵測 的聲源方位。

This thesis proposes a multiple sound sources localization and tracking method by microphone array. The method uses a multiple signal classification algorithm to detect the location of source and estimate their spectrum distribution. The multiple source tracking method is implemented by probability decision method combining the spatial and spectrum distribution. This proposed method can track multiple sources at the same time and maintain high detection rate under very low SNR conditions.