雙耳特徵差異分佈模版於非靜態聲源之定位研究

Abstract

在現實聲源定位的應用環境中，自然聲源之統計特性通常為非靜態（nonstationary），而環境則會造成複雜的迴響（reverberation）。因此，非靜態聲源於迴響環境中之定位，即成為工程學上重要的研究議題。本篇論文探討非靜態聲源與雙耳特徵差異（IPD、ILD）間之關係。在本篇論文中，採用移動極點模型的概念，提出以指數多項式建立非靜態聲源的強度波動模型。根據此模型，本論文提出利用IPD、ILD的分佈模版做為聲源定位之充分條件，並解釋分佈模版中多重峰值出現出原因。此外，本論文亦提出以高斯混合模型為基礎之「高斯雙耳特徵差異分佈模型」(GMBRDM)，作為非靜態聲源定位之演算法。此部分所提出之理論與演算法，皆有模擬或實驗結果加以討論與驗證。 除此之外，本論文將研究之非靜態聲源定位之方法應用於機器人室內定位環境，提出一創新之機器人位置與方向偵測系統。此系統適用於迴響複雜度高之環境，並具有對雜訊穩健之特性。實驗結果顯示，本系統可以用於近場與遠場環境，亦可在機器人與麥克風間無直接傳導路徑時使用。由於本系統可以執行機器人之全域定位，因此適合與其他定位方式整合，作為提供初始化參數或補償之用。

Nature sound sources are usually nonstationary and the real environment contains complex reverberations. Therefore, nonstationary sound source localization in a reverberant environment is an important research topic. This dissertation discusses the relationships between the nonstationarity of sound sources and the distribution patterns of interaural phase differences (IPDs) and interaural level differences (ILDs) based on short-term frequency analysis. The level fluctuation of nonstationary sound sources is modeled by the exponent of polynomials from the concept of moving pole model. According to this model, the sufficient condition for utilizing the distribution patterns of IPDs and ILDs to localize a nonstationary sound source is suggested and the phenomena of multiple peaks in the distribution pattern can be explained. Simulation is performed to verify the proposed analysis. Furthermore, a Gaussian-mixture binaural room distribution model (GMBRDM) is proposed to model distribution patterns of IPDs and ILDs for nonstationary sound source localization. The effectiveness and performance of the proposed GMBRDM are demonstrated by experimental results. The proposed nonstationary sound source localization algorithm is adopted for robot localization application. A novel and robust robot location and orientation detection method based on sound field features is proposed. Unlike conventional methods, the proposed method does not explicitly utilize the information of direct sound propagation path from sound source to microphones, nor attempt to suppress the reverberation and noise signals. Instead, the proposed method utilizes the sound field features obtained when the robot is at different location and orientation in an indoor environment. The experimental results show that the proposed method using only two microphones can detect robot’s location and orientation under both line-of-sight and non-line-of-sight cases and can be applied to both near-field and far-field conditions. Since this method can provide global location and orientation detection, it is suitable to fuse with other localization methods to provide initial conditions for reduction of the search effort, or to provide the compensation for localizing certain locations that cannot be detected using other localization methods.