The Role of Different Kernels in Classification of sEMG Signals for Automated Muscle Fatigue Detection Using SVM
Pysyvä osoite
Verkkojulkaisu
DOI
Tiivistelmä
Fatigue is a multidimensional and subjective concept, thus it is crucial to delineate the different levels and to quantify self- perceived fatigue. The aim of this study was to investigate the effect of different kernels on the accuracy of EMG signal classification into fatigue and nonfatigue stages. So, sEMG signals from right sternocleidomastoid muscle of nine healthy female subjects were recorded during neck flexion endurance test. Then six features in time, frequency and time- scale domains were extracted from the EMG signals. After intrinsic dimensionality estimation and reduction, linear and kernel SVM with polynomial, MLP and RBF kernels were used to classify feature vector. The results showed that the best accuracy (91/16%) is achieved via RBF kernel.