周期平稳信号盲源分离算法及其应用

BLIND SOURCE SEPARATION ALGORITHM OF SIGNALS ABOUT CYCLOSTATIONARY AND ITS APPLICATION

周期平稳信号盲源分离方法是在没有先验知识的前提下,从一组采集信号中提取未知源周期平稳信号。对混合的周期平稳信号进行盲源分离的算法进行研究。根据循环平稳度的特征,提出以循环平稳度作为盲源分离准则,取其旋转过程的极大值点,即为循环平稳度最大时,分离出周期平稳信号源。在理论推导的基础上,对分离准则进行证明。通过仿真验证,把该方法应用于齿轮箱振动数据实例分析中,表明循环平稳度作为分离准则有很好的分离未知源周期平稳信号的效果。

The method of blind source separation of signal about cyclostationary is to retrieve unknown source signals from a set of samples, without necessarily having any prior knowledge about the sources. Blind source separation of mixed signals about cyclostationary is studied. In the separation process, the degree of cyclostationary is proposed as the separation criterion. The separation criterion depends on the rotation variable only and the optimal rotation angle is maximizing the combined criterion. The algorithm of the separation criterion is validated. There lies in three respects. First, the criterion by a case is proved. Second, a set of simulated data is analyzed to validate it. Third, the signals which are recorded on the experimental test gearbox are used to validate the algorithm. The resuits demonstrate that the algorithm is useful for the analyzing of signals about cylcostationary.