Authors:
Tomasz Suchodolski
and
Andrzej Wolczowski
Affiliation:
Wroclaw University of Technology, Poland
Keyword(s):
Hand Prosthesis, EMG Signals, Decision Tree, Pattern Recognition.
Related
Ontology
Subjects/Areas/Topics:
Cybernetics
;
Health Engineering and Technology Applications
;
Human-Robots Interfaces
;
Informatics in Control, Automation and Robotics
;
NeuroSensing and Diagnosis
;
Neurotechnology, Electronics and Informatics
;
Robotics and Automation
Abstract:
The paper discusses the problem of the decision process of controlling the bio-prosthesis of the hand that is treated as the human intention recognition by means of the analysis of the electromyography (EMG) signals from the hand muscles. The number of movements, which is indispensable for the dexterity of the prosthesis, makes the recognition not entirely reliable. The approach presented herein includes three methods: the decision tree, neuron networks, and genetic algorithms in order to enhance the reliability of the EMG signal recognition. Simultaneously, the paper presents the software designed for the needs of the research and adapted to processing the EMG signals in compliance with these methods.