Do the generalized correlation methods improve time delay estimation of the muscle fiber conduction velocity?

Abstract

Muscle fiber conduction velocity is generally measured by the estimation of the time delay between electromyography recording channels. In the present paper, we propose to identify the best estimator of a constant time delay among those based on generalized correlation methods. To this end, small observation windows are considered and the fractional part of time delay was calculated using a parabolic interpolation. The results show that only Eckart and Hannan-Thomson approaches outperform the basic cross-correlation method when the signal to noise ratio (SNR) is up to 0 dB and the observation duration is 250 ms. This study will be a background for further extension to time-varying delay estimation.