Time-Frequency Distribution of SEMG Pattern Recognition in Reducing Limb Position Invariant

摘要

According to the World Health Organization, about 160,000 people in Malaysia are required to use prostheses. One of the factors that affect the current prosthesis control is that the variation in the limb position and normal use results in different electromyogram (EMG) signals with the same movement at different positions. Thus, the goal of this study is to make sure that amputees can control their prosthetics in an accurate manner regardless of its hand movement and limb position. This paper uses time-frequency distribution to extract the EMG feature and six SVM classification learners; linear SVM, quadratic SVM, cubic SVM, fine Gaussian SVM, medium Gaussian SVM, and coarse Gaussian SVM were compared to find the most suitable one for this application. The performance of the analysis is then verified based on classification accuracy. From the results, the overall accuracy for the classification is 65% (linear SVM), 87.5% (quadratic SVM) and 97.5% (cubic SVM), 100% (fine Gaussian SVM), 70% (medium Gaussian SVM), and 45% (coarse Gaussian SVM), respectively. It is hoped that the study could serve as an insight to improve conventional prosthetic control strategies.