基于半球面共形阵列的点声源辨识方法

The method of point-acoustic source identification based on hemispherical conformal array

采用20个全向声压传感器作为对空间声源信号的检测单元,按C60结构非均匀分布形成半球面共形阵列,提出多级分离GDE改进方法,并对其调整因子进行参数优化。实验结果表明,信噪比在-10dB^10dB范围内,当声源方位和频率都不同时,改进GDE方法准确识别二声源的信噪比区间为-7dB^10dB,相较传统方法将SNR在-10dB^-7dB内的识别准确率最低提高到80%;SNR在-10dB^0dB内,三声源识别准确率相较传统方法提高40%~60%;四、五、六声源识别准确率比传统方法平均提高50%~80%。

20 omnidirectional sound pressure sensors are used as the detection units for spatial acoustic source,which is formed a hemispherical conformal array referring to the non-uniformly distributed C60.Proposed improved multi-stage separating GDE,in addition,the adjustment factors are optimized.The experimental results show that the signal-to-noise ratio is in the range of-10dB^10dB,under the different direction and frequency of source,the sound source azimuth and frequency are not at the same time,the improved GDE method can accurately identify the signal-to-noise ratio of the two sound sources in the range of-7dB^10dB,which improves the lowest recognition accuracy rate of SNR within-10dB^-7dB is increased to 80%compared with the traditional method,the accuracy rate of three-sources recognition is higher than that of the traditional method about 40%~60%which SNR within-10dB^0dB,the accuracy rate of four-sources,five-sources and six-sources recognition is advanced about 50%~80%averagely.