Improved algorithm of muscle fatigue detection using linear regression analysis

Conventional linear regression analysis methods used to measure muscle fatigue based on mean power frequency (MPF) can yield results that are dependent on the epoch of each MPF value. An improved algorithm containing two window segmenting methods, overlapped and non-overlapped schemes, is provided to investigate the effect of window length on linear regression analysis during sustained isometric constant force muscle contraction. The minimal length of each segment can be determined by making the slope sign changes to be zero. In conclusion, the minimal length is independent of the epoch of each MPF value.