黄信圣*,丁浩**,黄琦***,高增梁*.基于交叉谱方法的三维波束形成声源定位性能研究[J].高技术通讯(中文),2024,34(12):1286~1296 |
基于交叉谱方法的三维波束形成声源定位性能研究 |
Research on location performance of three-dimensional beamforming sound source based on cross spectrum method |
|
DOI:10. 3772 / j. issn. 1002-0470. 2024. 12. 005 |
中文关键词: 声学; 阵列信号处理; 声源定位; 麦克风阵列; CSM-B; CSM-EF |
英文关键词: acoustics, array signal processing, sound source localization, microphone array, CSM-B, CSM-EF |
基金项目: |
作者 | 单位 | 黄信圣* | (*浙江工业大学化工过程机械研究所杭州 310023)
(**浙江省机电设计研究院杭州 310051)
(***东南大学机械工程学院南京 211189) | 丁浩** | | 黄琦*** | | 高增梁* | |
|
摘要点击次数: 20 |
全文下载次数: 58 |
中文摘要: |
近年来声学麦克风阵列及相关的信号处理技术被广泛用于声源定位。其中,基于波束形成的声学成像方法(BBAI)在二维平面阵列上实现了对噪声源的三维成像,并可识别噪声源深度。然而,基于平面麦克风阵列,客观地发现了在阵列深度方向上出现了完全相似的伪声源。此外,基于传统波束形成算法的三维声源成像图中有明显的主瓣宽度大、空间分辨率较低(特别是深度方向上)的问题。本文在基于波束形成的声学成像方法上,分别在BBAI基础上提出了2种结合交叉谱方法(CSM)的波束形成声学成像算法——CSM-B和CSM-EF,并通过三维锥型阵列(CSA)和二维轮辐阵列(SWA)比较了CSM-B和CSM-EF在主瓣宽度和总运行时间上的性能。通过对单极子声源的测试,对3种算法进行了理论评估,并在全消音室中进行了实验评估。结果表明,从阵列角度上看,相较于二维轮辐阵列,三维锥型阵列能有效地减小主瓣宽度,尤其是在特定的声源深度范围内,效果较为明显。此外,从3种算法的角度上看,CSM-EF以牺牲较大的计算时间为代价,有效解决了主瓣宽度大、旁瓣数量多的声源定位问题,提供了最好的三维声源成像图,效果最佳。 |
英文摘要: |
In recent years, acoustic microphone array and its related signal processing technology have been widely used in sound source localization. Among them, the beamforming-based acoustic imaging method (BBAI) can realize the three-dimensional imaging of the noise source on the two-dimensional planar array, and identify the depth of the noise source. However, based on the planar microphone array, it is objectively found that the pseudosound source appears exactly similar in the depth direction of the array. In addition, there are obvious problems of large mainlobe width and low spatial resolution (especially in the depth direction) in three-dimensional sound source imaging based on traditional beamforming algorithm. In this paper, two beamforming-based acoustic imaging algorithms——beamforming-based acoustic imaging method based cross spectrum method (CSM-B), and error function based cross spectrum method (CSM-EF) combined with cross spectrum method are proposed based on BBAI, and the performance of CSM-B and CSM-EF in mainlobe width and total running time is compared by three-dimensional cone shape array (CSA) and two-dimensional slice wheel array (SWA). The three algorithms are evaluated theoretically by testing the monopole sound source and experimentally in a full anechoic chamber. The results show that compared with the two-dimensional slice wheel array, the three-dimensional cone shape array can effectively reduce the width of the main lobe, especially in the specific depth range of the sound source. In addition, from the perspective of the three algorithms, CSM-EF can effectively solve the problem of sound source location with large mainlobe width and large number of side lobes at the cost of large calculation time, and provide the best three dimensional sound source imaging map with the best effect. |
查看全文
查看/发表评论 下载PDF阅读器 |
关闭 |
|
|
|