Table 6.

Five compared feature extraction methods and the corresponding operations.

Name	Operation
Mean	X Mean=(1/N)∑n=1Nx(n)
RMS	$X_{\text{RMS}}=\sqrt{\left(1/N\right){\sum }_{n=1}^N{x}^2\left(n\right)}$
EMD	Empirical mode decomposition (EMD) is a method of signal decomposition based on the time-scale characteristics of the data itself, the detailed process of which can refer to [46].
DWT	Discrete wavelet transform (DWT) is a discrete wavelet transform method. Its detailed process can refer to [47].
ASR	Artifact subspace reconstruction (ASR) is relatively new technique, and it is based on new approach of signal reconstruction with the reference signal fragment. The detailed process of ASR can refer to [48].