论文部分内容阅读
目前使用的大多数盲源分离方法都依赖于观测传感器数量等于信号源数目这样一个基本假设。主要针对传感器数量大于源信号数量情况,对涡扇发动机振动信号盲源分离问题进行研究,先后采用三种算法进行试验,结果证明了奇异值降维方法和基于四阶累积量的源数估计方法的可行性和可靠性。
Most current methods of blind source separation rely on the basic assumption that the number of observed sensors equals the number of sources. In this paper, the problem of blind source separation of turbofan engine vibration signal is mainly studied in the case that the number of sensors is larger than the number of source signals. Three algorithms are used successively to test. The results show that the singular value reduction method and source estimation method based on fourth-order cumulants Feasibility and reliability.