基于独立分量的信号分离算法研究

Research on Signal Separation Algorithm Based on Independent Component

噪声控制工程中,定量分离主要噪声源是一个非常复杂的问题,国内外对此进行了大量研究,并提出了许多算法,但基本都只适用于线谱,对于宽带信号显得无能为力。据此,基于独立分量分解原理,尝试通过信号分解、线谱提取的方式,将宽带噪声从原信号中准确分离出来,然后可以依据相干分析和谱分析等手段定量分析。计算机仿真研究效果良好,证明该方法具有一定工程应用价值。

In the noise control engineering, the quantitative separation of the main sources is a very complex problem. There are many methods to be used as sources identification, but the majority methods are fit for line spectrum, and they are not fit for the identification of broadband noise. For solving the problem, the method of separation broadband noise from complex noise is presented, which is based on the independent component a- nalysis and the mode of trying to decompose signals, abstract line spectrum ware are also given. It can be ap- proved that the effectiveness and value in the project of the noise and vibration control.