Comparison of Two Muscle Activity Detection Techniques from Surface EMG Signals Applied to Countermovement Jump

The goal of this research was to compare two methods for automatic muscle activity detection from surface EMG signals: an algorithm based on wavelet transform and a simple threshold based algorithm (low- and high-pass filtering and comparison with a threshold). A total of 120 recorded surface EMG signals, obtained by measurements on 12 different leg muscles of professional handball players performing countermovement jump, were analyzed using both algorithms. An experienced researcher visually determined onset/offset times and these were used as a reference. For each signal bias between evaluated algorithms and visual evaluation of muscle activity was calculated. The results indicate that the algorithm based on wavelets is more appropriate for the analysis of fast dynamic movements. Low level of background noise presented a problem whit establishment of an appropriate activity threshold. For the wavelet based algorithm this was dealt with by adding low level of white noise to the raw EMG signals. This procedure improved the validity of the activity threshold defined for EMG signals with low levels of background noise therefore improving the results of the algorithm. However, it does not solve the issue of inappropriate threshold when using the simple threshold based algorithm.