论文部分内容阅读
为适应汽车加速噪声非平稳工作过程声源识别的需要,提出了一种基于连续小波变换的时频相干估计方法。应用该方法从时频空间揭示了汽车加速时,标准测点噪声信号与发动机近场噪声信号的关系。与基于FFT的估计方法对比表明,该方法能够较好地识别局部时间段、低信噪比条件下2个噪声信号的关联程度。
In order to meet the need of sound source identification during non-stationary car acceleration noise, a time-frequency coherent estimation method based on continuous wavelet transform is proposed. This method is used to reveal the relationship between the noise signal of the standard measuring point and the near-field noise of the engine when the vehicle accelerates from the time-frequency space. The comparison with the FFT-based estimation method shows that this method can better identify the correlation degree of two noise signals under the condition of local time and low signal-to-noise ratio.