Periodic autoregressive (PAR) time series is considered as one of the most common models of second-order cyclostationary processes. In real applications, the signals with periodic characteristics may be disturbed by additional noise related to measurement device disturbances or to other external sources. The known estimation techniques for PAR models assume noise-free model, thus may be inefficient for such cases. In this paper, we propose four estimation techniques for the noise-corrupted finite-variance PAR models. The methodology is based on Yule-Walker equations utilizing the autocovariance function. Thus, it can be used for any type of the finite-variance additive noise. The presented simulation study clearly indicates the efficiency of the proposed techniques, also for extreme case, when the additive noise is a sum of the Gaussian additive noise and additive outliers. This situation corresponds to the real applications related to condition monitoring area which is a main motivation for the presented research.
翻译:定期自动递减(PAR)时间序列被认为是第二序周期周期性静止过程最常见的模型之一。在实际应用中,定期特性的信号可能会被与测量装置扰动或其他外部来源有关的额外噪音干扰。已知的PAR模型估计技术假定无噪音模型,因此对此类情况来说可能效率低下。在本文件中,我们建议对噪音干扰的有限变异PAR模型采用四种估计技术。该方法基于利用自动变换功能的Yule-Walker方程式。因此,该方法可用于任何类型的有限变异添加噪音。所介绍的模拟研究清楚地表明了拟议技术的效率,在极端情况下,当添加噪音是高斯添加噪音和添加异物外体的总和时,这种情形与条件监测领域的实际应用相对应,而这是提出研究的主要动机。