基于声源预估的波叠加组合全息声源识别方法  

A hybrid sound source identification method with wave superposition algorithm and acoustic holography based on source pre-estimation

作　　者：李兵[1] 杨殿阁[1] 郑四发[1] 李克强[1] 连小珉[1] 

机构地区：[1]清华大学汽车安全与节能国家重点实验室,北京100084

出　　处：《声学学报》2010年第6期646-652,共7页Acta Acustica

基　　金：国家自然科学基金资助项目(50605038)

摘　　要：为解决声源识别中旁瓣效应等因素引起的虚假识别结果问题,提出了基于声源位置预估的波叠加组合全息算法。应用波束形成算法进行声源位置的预估,将预估声压峰值点作为假定声源点,通过波叠加原理建立方程组,利用声源强度求解结果构造完整全息平面,通过声全息方法重新计算重建面声压。进行了仿真和实验验证,经该算法重构后,波束形成算法产生的虚假声源处的声压得以明显降低。表明该算法可以抑制旁瓣效应等因素的影响,提高声源识别的准确率。To solve the problem of ghost sources in the noise source identification,a hybrid method with the wave superposition algorithm and acoustic holography based on pre-estimating the sources＇ positions was proposed.The sources＇ positions are pre-estimated with beam-forming algorithm,and source strengths at the pre-estimated positions are figured out with the wave superposition method.Then,the sound pressure of the full holography plane is constructed with the source strengths,and the sound pressure of the reconstruction plane is re-computed with the acoustic holography method.The simulations and experiments were carried out,the sound pressures of the ghost sources produced by the beam-forming algorithm were reduced greatly with the proposed method.The results indicate that the side lobe effect can be restrained,and the sources can be identified more correctly with this method.

关 键 词：声源识别 组合全息 识别方法 波叠加 波束形成算法 叠加原理 峰值点 声压 

分 类 号：O421.2[理学—声学]