Classification of Surface EMGs Using Wavelet Packet Energy Analysis and a Genetic Algorithm-Based Support Vector Machine

Abstract

The aim of our study was to recognize results of surface electromyography (sEMG) recorded under conditions of a maximum voluntary contraction (MVС) and fatigue states using wavelet packet transform and energy analysis. The sEMG signals were recorded in 10 young men from the right upper limb with a handgrip. sEMG signals were decomposed by wavelet packet transform, and the corresponding energies of certain frequencies were normalized as feature vectors. A back-propagation neural network, a support vector machine (SVM), and a genetic algorithm-based SVM (GA-SVM) worked as classifiers to distinguish muscle states. The results showed that muscle fatigue and MVC could be identified by level-4 wavelet packet transform and GA-SVM more accurately than when using other approaches. The classification correct rate reached 97.3% with seven fold cross-validation. The proposed method can be used to adequately reflect the muscle activity.