Özet:

In this study, the effects of the classification of extender and flexor tendons, along with the classification of the hand opening/closing movement using the electromyogram (EMG) signs, were studied. With the advanced technology, the need for smart prothesis hands is rapidly increasing. For this purpose, the EMG signs, measured from a total of 32 persons by surface electrots from the front and rear parts of the front arm, were analyzed by time frequency techniques and the classification of the open/cover movement was carried out by different classifiers. It is also compared by showing the achievements on the classification of EMG signs obtained from each channel and at the same time. The use of separate wave conversion (ADD) for the analysis of non-stop signs such as EMG, where the short-term Fourier conversion remains insufficient, is a good solution. Signals divided into the subband with ADD provide high time resolution in high frequencies, low frequency resolution and high frequency resolution in low frequencies, low time resolution. The results of the ADD analysis were obtained with specialties, and these characteristics were used for the training and testing of different classifiers. All the data were used in both training and testing stages of classifiers with 10fold cross-verification technique. The best outcome from the classifiers used was 97.25% with overlearning machines (ELM). This study is expected to contribute to the development of smart prothesis handmade, along with the up-to-date technology in the biomedical field.

Anahtar Kelimeler: