Abstract
Amputees around the world barely have any access to top-notch, smarter prosthetics due to the fact that they are either inaccurate or cost-inefficient. One of the more challenging tasks is the accurate detection of gestures and this paper presents a new assembly of segmentation, discrete wavelet decomposition, subbands features extraction, and machine learning algorithms for an automated identification of hand gestures by processing the Surface Electromyogram (sEMG) signals. A comparative analysis of two machine learning algorithms namely, Decision Tree and K-Nearest Neighbor is performed for processing the designed features set. Results confirm that the performance of proposed method is comparable or superior to the counterparts. The applicability is tested by using a publicly available sEMG dataset. Six different hand gestures for a mono male subject are considered. The system secures a highest classification accuracy of 93% for the case of Decision Tree algorithm.