Separation of stationary and non-stationary sources with a generalized eigenvalue problem

EOG Artifact Correction from EEG Recording Using Stationary Subspace Analysis and Empirical Mode Decomposition

“…The method utilizes stationary subspace analysis (SSA) [28,29] in conjunction with empirical model decomposition (EMD) [30]. Unlike the classic blind source separation with ICA or SOBI, the adopted BSS algorithm SSA is explicitly tailored to the understanding of distribution changes [29]. The type of distribution changes that SSA detects are changes in both the mean and the covariance matrix.…”

Fault Diagnosis of Wind Turbine Gearbox Based on Wavelet Transform

“…The amplification of resonance is used to extract these high impact frequencies by envelope spectroscopy. In this paper, the characteristics of fault signals in the background of strong noise interference are realized, and the purpose of fault diagnosis is realized [11,12] . Taking into account the above issues, wavelet analysis has the characteristics of multi-scale analysis of signals, which can be combined with the two methods, namely, using wavelet transform to perform signal processing on multi-frequency scales, and applying envelope decomposition to each scale in wavelet analysis techniques.…”

Independent Component Analysis

“…The dimensionality of the data can be reduced to the most nonstationary directions, which are most informative for detecting state changes in the time series. Analytic stationary subspace analysis [41] solve a generalized eigenvalue problem. The solution is guaranteed to be optimal under the assumption that the covariance between stationary and nonstationary sources is time-constant.…”