基于压缩感知理论的运动声源识别方法研究

摘要

In view of the complexity of the traditional signal processing of acoustic source identification,a method of applying compressive sensing theory to the research of sound source identification is presented.According to the characteristic of the moving sound source,the best method of the compression and reconstruction of the moving sound field is the Discrete Fourier Transform (DFT).Through the contrast analysis of three different test methods of sound field reconstruction movement,namely,acoustic source identification of acoustic phase instrument, LMS array recognition and LMS acquisition data for compression and reconstruction identification,besides,to verify the feasibility of the proposed method. It provides an instructive view of the identification of the high speed,variable speed and multi-point sound source in the future.%针对于传统的声源识别信号处理过程繁琐复杂问题,提出了一种将压缩感知理论运用到声源识别研究中的方法.根据运动声源的特性研究给出了运动声场的压缩和重建的最佳方法是离散傅里叶变换(DFT)时变信号稀疏化和特征基匹配追踪的方法.通过实验对比分析了运动声场重建的三种不同测试方法即声像仪声源识别、LMS阵列识别以及LMS采集数据进行压缩重构识别,验证了提出方法的可行性.为将来的高速、变速、多点声源的识别研究提供了指导意义.