“…The PSD of a random signal can be estimated using classical methods (periodogram or Blackman-Tukey estimators) or modern parametric model methods (autoregressive (AR), moving average (MA), autoregressive moving average (ARMA)) [1,12]. However, PSD estimation unavoidably introduces a number of factors (method for PSD estimation, implementation algorithm, order of parametric model, shape and size of the analysis window) that directly affect the estimates of the spectral variables [9,10,15,35,38]. Furthermore, the derivation of meaningful, statistically-significant spectral parameters requires some assumptions regarding the characteristics of the signal [13].…”