sEMG FEATURE EXTRACTION USING HYBRID TECHNIQUES FOR POWER X.A

This research is about the surface Electromyography (sEMG) feature extraction using hybrid method for Powered Exoskeleton Arm (Power X.A) application. The main objective of this research is to investigate the feature extraction techniques for EMG signal processing. This report is divided into 5 chapters. The first chapter is about the introduction, the second chapter is on the literature review and theory of this research, the third chapter is on the methodology used in this project, the fourth chapter is the discussion of the results and the final chapter is the conclusion and recommendation of this research. EMG is the biomedical signal and widely in used in clinical applications. This research can be divided into 3 parts where the 1st part is on the design on the experimental procedure, the 2nd part is on the signal acquisition and the 3rd part is on the feature extraction based on hybrid techniques. The raw EMG signal was collected from different test subjects and further processed in MATLAB to obtain the clean EMG signal. The most powerful EMG feature extraction which is wavelet techniques and mean absolute value was used for this research. The result shows that Daubechies wavelet order 7 in level 1 and 2 gives the best performance in EMG feature extraction.