Defining the Features of EMG Signals on the Forearm of the Hand Using SVM, RF, k-NN Classification Algorithms
Adilbek TurgunovInformation technology, Karshi branch of the Tashkent University of Information Technologies named after Muhammad Al-Khwarizmi, Karshi, UzbekistanKudratjon ZohirovComputer Systems, Tashkent University of Information Technologies named after Muhammad Al-Khwarizmi, Tashkent, UzbekistanAlisher GaniyevInformation technology, Karshi branch of the Tashkent University of Information Technologies named after Muhammad Al-Khwarizmi, Karshi, UzbekistanBarno SharopovaInformation technology, Karshi branch of the Tashkent University of Information Technologies named after Muhammad Al-Khwarizmi, Karshi, Uzbekistan
2020en
ABI
Аннотация
EMG (electromyogram) signal-based control systems that provide important information on muscle activity are now in development. In this article, through some of the features of the EMG signal, three hand movements are classified. Experiments show that by using with the combination of Simple Square Integral (SSI), Average Amplitude Change (ACC), Integrated EMG (IEMG), Waveform Length (WL), Root Mean Square (RMS), Mean Absolute Value (MAV), Zero Crossing (ZC), Wilson Amplitude (WAMP) and Log-Detect (log-D) of EMG signal, 99% result of the process of signal recognition used in the classifier algorithms in k-NN (K -nearest neighbor) and RF (Random Forest), 96% result in SVM (Support vector machine) is achieved.
Перевод пока недоступен