Feature Extraction Of Forearm EMG Signal For Exoskeleton Hand

Tarmizi, Aine Ilina (2015) Feature Extraction Of Forearm EMG Signal For Exoskeleton Hand. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text Feature Extraction of Forearm EMG Signal for Exoskeleton Hand - 24 pages.pdf - Submitted Version Download (655kB)

Abstract

Electromygraphy (EMG) signal is non-stationary, non linear, and have large variation in its signal. It consists of different type of noises in the EMG signal, some of the noises can be eliminated by filtering. However, filtering cannot eliminate random noises of the EMG signals. Therefore, feature extraction has to be performed to the EMG signals to eliminate the unwanted parts of the EMG signals. The first objective of this project includes the forearm raw EMG signals extraction for exoskeleton hand. The next objective is to perform feature extraction on the raw EMG signals that are extracted in objective one. Lastly, the feature extraction performance is analyzed by using the percentage error calculation approach. The method of this research is that, the experiment on extracting the EMG signals are done by using the Muscle Sensor V3 Kit. The performance and analysis of the feature extraction are done by using the MATLAB software. Literature review covers the theory and basic principles and review of previous work on EMG signals, muscle selection, signal conditioning of DAQ, and feature extractions. The movement of hand close (HC), hand open (HO), wrist flexion (WF) and wrist extension (WE) is selected for this research. The selected muscle are FDS, FCR and ECRL muscles. The feature extraction of IEMG and MAV is performed to the EMG signals. The expected result of this research will be the raw EMG signal, feature extracted EMG signal, and also the analysis of the performance. It is expected that the better performance will have the lower percentage of error.

Actions (login required)

View Item