气体泄漏声源的时域压缩波达方向估计研究

DOA estimation of a gas leakage sound source based on analog signals’ time-domain compression sensing

针对现有压缩感知算法中对模拟信号压缩采样的不足以及传统测量矩阵存储量和计算量大等问题,提出一种基于模拟信号时域压缩的波达方向(DOA,Direction Of Arrival)估计算法,并应用于气体泄漏声源方位估计研究。首先建立气体泄漏声源DOA估计模型,然后设计一种直接非均匀随机欠采样方案,并构造相对应的形式简单的等效测量矩阵,实现对模拟声源信号的直接压缩采样,避免了传统测量矩阵数据量大及计算复杂度高等问题,最后采用子空间追踪算法进行重构,实现了较快的计算速度和较高的重构精度。理论分析和实验表明,该算法能成功实现气体泄漏声源的DOA估计,且在显著降低信号采样率和计算复杂度的同时,实现了较高的估计精度和分辨率,估计性能更佳。

Aiming atproblems ofshortage of analog signal compression sampling in existing compression sensing algorithmsand large amount of memory and calculation of traditional measurement matrices,a direction of arrival(DOA)estimation algorithm based on analog signals’time-domain compression was proposed and applied to study DOA estimation of a gas leakage sound source.Firstly,a DOA estimation model of a gas leakage sound source was established.Then a direct non-uniform random under-sampling scheme was designed and the corresponding equivalent measurement matrix with a simple formwas constructed to realize the direct compression sampling of analog sound source signals,and avoidlarge amount of data and high computational complexity of traditional measurement matrix.Finally,the subspace tracking algorithm was adopted to reconstruct acoustic signals,and realize faster calculation speed and higher reconstruction accuracy.Theoretical analysis and tests showed that the proposed algorithm can successfully realizethe DOA estimation of a gas leakage sound source;it can significantly reduce sampling rate and computational complexity,and realize higher estimation precision and resolution,so it has a better estimation performance.