摘要
反卷积DAMAS2波束形成方法能有效消除旁瓣、提高分辨率和阵列的动力学水平。基于数值模拟的声源识别成像图及标准差曲线探究其结果随声源类型、信号频率、迭代次数的变化规律,结果表明:反卷积DAMAS2波束形成在识别单声源、不相干声源、相干声源时均可有效消除旁瓣、提高分辨率;频率越高,迭代次数越多,分辨率越好;声源识别准确度随迭代次数的增加呈"L"形变化,转折点对应迭代次数约为100;综合考虑计算准确度和效率,推荐100为优选迭代次数。上述结论对反卷积DAMAS2波束形成技术的运用具有指导意义。进一步,基于开发的声源识别软件进行的单扬声器、不相干双扬声器、相干双扬声器的声源识别试验验证了该方法的有效性及开发软件的正确性。
（2 Deconvolution approach for the mapping of acoustic sources, DAMAS2 ） beamforming sound source identification method could not only eliminate sidelobes but also improve resolution and dynamic level of the array. The varying rules of the sound source identification results with source type, signal frequency and iteration number were explored based on numerical simulated sound source identification imaging maps and standard deviation curves. The results are listed as following. Firstly, the DAMAS2 beamforming method could eliminate sidelobes and improve resolution effectively in identifying single source, incoherent and coherent sources. Secondly, the higher the frequency and the more the iteration number, the better the resolution. Thirdly, the curve of the sound source identification accu- racy versus iteration number shows an ＂L＂ shape variation and the iteration number corresponding to the turning point is about 100. Finally, 100 iterations are recommended taking both accuracy and efficiency into account. These conclusions have guiding significance for the application of the DAMAS2 beamforming method. Furthermore, the sound source identification software was developed and the sound source identification experiments on a single loud- speaker, two incoherent loudspeakers and two coherent loudspeakers were carried out, which verifies the effectiveness of this method and the correctness of the developed software.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2013年第8期1779-1786,共8页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(50975296)
牵引动力国家重点实验室开发课题(TPL0903)资助项目
关键词
反卷积
波束形成
声源识别算法
软件开发
试验
deconvolution
beamforming
sound source identification algorithm
software development
experiment
