Fast on-line polarity correlation algorithms for muscle fibre conduction velocity measurement

Polarity cross-correlation is a useful technique for the measurement of muscle fibre conduction velocity using surface electromyography. Owing to the nature and volume of computation involved in the correlation function, standard techniques for its estimation by a microprocessor are too slow for an on-line application. In this paper two algorithms suitable for on-line estimation of polarity function are presented. Some useful features of the correlation function, as well as careful programming and careful choice of instructions, made it possible to use a standard microprocessor to achieve higher sampling rates than those reported recently.