DEVELOPMENT OF MUSCLE-COMPUTER INTERFACE SYSTEM FOR WRIST MOTION CLASSIFICATION

• 主要作者: Rakhmadi, Imam
• 格式: Final Project
• 語言: Indonesia
• 在線閱讀: https://digilib.itb.ac.id/gdl/view/81581
• 機構: Institut Teknologi Bandung

The wrist is an essential part of the body for humans to perform various activities. The basic movements that occur in the wrist commonly involve flexion, extension, and a straight state between flexion and extension (normal). These motions can be identified by surface electromyography (sEMG) signals. The existence of this sEMG signal method allows humans to communicate or interact with computers through their body's muscle activity. However, the acquisition, signal processing, and motion classification of these sEMG signals are very complex. Hence, this research aims to design a sEMG signal acquisition and signal processing method for wrist motion, as well as design a classification algorithm to identify wrist motion including flexion, extension, and normal motion. This research is the initial stage of developing a muscle-computer interface system for wrist motion. The major steps in developing the muscle-computer interface of this research include data acquisition using the Cyton Board OpenBCI, signal processing, and motion classification with the support vector machine (SVM) algorithm which was performed offline and not in real time. The features used as SVM input are root mean square (RMS) features with a moving window. This research has obtained a good design for sEMG signal acquisition and signal processing so that RMS features can be extracted for classification purposes. In addition, the algorithm design for motion classification performed quite well with 90.27% accuracy, 90.61% precision, 90.17% sensitivity, and 90.27% f1-score. This shows that the muscle-computer interface has been able to classify wrist movements, including flexion, extension, and normal with a fairly high level of accuracy. Keywords: sEMG, Muscle Computer Interface, RMS, SVM