基于快速迭代柯西阈值的声场重建方法研究  

Sound field reconstruction method based on a fast iterative Cauchy thresholding algorithm

作　　者：黄琳森 徐中明[1] 昝鸣 张志飞[1] 贺岩松[1] HUANG Linsen;XU Zhongming;ZAN Ming;ZHANG Zhifei;HE Yansong(College of Mechanical and Vehicle Engineering,Chongqing University,Chongqing 400044,China)

机构地区：[1]重庆大学机械与运载工程学院,重庆400044

出　　处：《振动与冲击》2023年第6期272-279,共8页Journal of Vibration and Shock

基　　金：国家自然科学基金(11874096)。

摘　　要：在声源识别领域,稳定且准确识别目标声源是一项极具挑战性的任务,尤其在中、低频范围内,因此论文提出快速迭代柯西阈值算法在中、低频段实现鲁棒性较强的声源识别和定位。通过蒙特卡罗分析,与3种经典罚函数方法Tikhonov正则化方法、宽带声全息算法和内点法求解方法相比,所提出的方法在中、低频段获得了稳定的声压重建结果,平均重建误差在5%以下,对声源识别逆问题具有良好的距离适应性和信噪比适应性。In the field of sound source identification,it is a challenging task to identify the target sound source stably and accurately,especially in the middle and low-frequency ranges.A fast iterative Cauchy threshold algorithm(FICTA)was proposed to identify the sound source and achieve robust sound source location in the middle and low-frequency ranges.Compared with the other three classical penalty function methods(including Tikhonov regularization method,wideband acoustic holography algorithm and interior point method),the Monte Carlo analysis shows that the proposed method could obtain the stable sound pressure reconstruction results whose average reconstruction error is below 5%in the medium and low-frequency ranges,and it has favorable distance-adaptability and signal-to-noise ratio-adaptability to the inverse problem of the sound source identification.

关 键 词：柯西阈值 麦克风阵列 近场声全息 声场重建 声源识别 

分 类 号：TH73[机械工程—仪器科学与技术]