基于声压幅度比的多声源分离定位决策研究

摘要

The sound pressure amplitude ratio model was applied to the study of multiple sound source localization, and a method for multi-sound source separation and localization based on the sound pressure amplitude ratio was proposed.The method uses the blind signal separation algorithm to separate mixed signals and uses the spectrum estimation similarity method to determine the distribution of each sound source in the mixed signals.Then, according to the amplitude difference factor, the method obtains the distribution of sound source signals of the sensor and uses the sound pressure amplitude ratio model for single sound sources to determine sound source positions to achieve the multi-sound source localization.The blind signal separation algorithm is relatively mature, and the actual sound source signals are mostly non-Gauss, so it satisfies the condition of blind signal separation.The proposed method is practical, and can be widely used to solve the similar problems in signal localization and separation fields.%利用声压幅度比模型,提出了一种基于声压幅度比的多声源分离定位方法,该方法利用盲信号分离算法实现混合声源信号的分离,根据谱估计的相似度确定接收信号中各声源的分配情况,结合幅度差异因子获得传感器的声源信号分布,再通过单声源的声压幅度比模型确定声源位置,实现多声源定位.由于盲信号分离算法比较成熟,且实际中的声源信号大多为非高斯,因而满足盲信号分离条件.该方法具有实用强、应用性广等特点,对其它分离、定位问题也有借鉴作用.