Abstract
The hands play a significant role for grasping and manipulating different objects. The loss of even a single hand affect the human activity and the prosthetic hand is a solution to equip the armless subject. Recognition of the surface electromyogram (sEMG) patterns is vital in the design of prosthesis hand. For this reason, there is a need for a proper analysis of muscle behavior. In this research, sEMG signals are used to control a robotic hand by using discrete wavelet transform (DWT) and bagging. Different types of muscle contraction can cause EMG signals to vary, affecting classification performance. In this study, MSPCA is used for denoising and DWT is used for feature extraction to evaluate their efficiency at classifying sEMG signals, which were recorded during the grasping movements with various objects. Furthermore, the performance of different classifiers with bagging were studied. An effective combination of DWT, and Bagging achieves good performance, using k-fold cross validation regarding to the total classification accuracy. The proposed methods in this study have potential applications in the prosthetic hand control.