| Summary: | 碩士 === 國立高雄應用科技大學 === 電機工程系博碩士班 === 102 === In this thesis, we use a microphone array with blind source separation algorithm and time difference of arrival (TDOA) in far-situation to do multiple sound source localization and source number estimation. Most of the researches of multiple sound source localization are focused on the angle of source. However, there is a problem that the user cannot identify the corresponding angle of each sound source, and we name this problem as data association. To solve this problem, we use nonnegative matrix factor 2-D deconvolution (NMF2D) to separate the mixed single and then we can consider the multiple sound source localization as many independent sound sources localization. Therefore, we can calculate the corresponding angle of each sound source. We also use the interaural level difference (ILD) and TDOA to estimate the distance of each sound source. Finally, the multiple sound source localization system is programmed by LabVIEW 2012 software.
|
|---|
