Table 9.

Pre-processing and feature extraction for vibration approach.

Authors	Processing Techniques	Feature Extraction
Bortnowski et al. [44]	Filtering and normalization using the mean.	Spectrogram, signal autocorrelation (ACF), mean peak frequency.
Li et al. [4]	WPD.	The energy of each frequency band is extracted as the feature
Wijaya et al. [12]	Fast Fourier transform.	Envelope analysis.
Ravikumar et al. [45,46,47,48]	Trimmed off to ensure the uniform length of the signal.	The various parameters are mean, median, mode, standard error, standard deviation, kurtosis, skewness, minimum value, maximum value, sample variance, range.
Wijaya et al. [60]	Wavelet transform.	Extracted features from WT and statistical time-domain features (e.g., RMS, peak-to-peak, and standard deviation)
Muralidharan et al. [58]	Trimmed off to ensure the uniform length of the signal.	The various parameters are mean, median, mode, standard error, standard deviation, kurtosis, skewness, minimum value, maximum value, sample variance, range.
Roos and Heyns [23]	WPD.	A sum of squares of the frequency magnitudes of each wavelet.