Classification of multifunction surface EMG using advanced AR model representations
R. Knox, Dana H. Brooks · 2002
The potential use of new features derived from autoregressive coefficients for upper limb EMG pattern recognition has been discussed (Knox et al., 15/sup th/ Ann. Int. Conf. of the IEEE EMBS, San Diego, CA, USA, Oct. 1993). These new features (reflection coefficients, cepstral coefficients, and logarithmic area ratios) are used heavily in speech processing. Results from a nonparametric linear classifier are presented.>