1Research scholar, Karpagam University, Coimbatore, India
2Associate Professor, Faculty of Engineering, Karpagam University, Coimbatore, India
3Research Scholar, Karpagam University, Coimbatore, India
Online published on 21 September, 2018.
This paper evaluates time domain, frequency domain and time scale features to classify upper limb motions using myoelectric signals (MES). A comparative study is made among the features to decide the suitability in myoelectric prosthetic control. The system was trained using MES corresponding to six upper limb movements at three different force levels. Time domain features outperformed other two features when used with LDA classifier. However time scale features could exhibit supremacy when classified using an SVM classifier. PNN classifier maintained a fairly good classification accuracy for all the three features.
Electromyogram, myoelectric prosthesis, pattern recognition, contraction level variation, robust control, classification accuracy, transradial
