Finger and hand movement recognition and classification aiming at hand prosthesis control

• Grant number: 13/12897-5
• Support Opportunities: Scholarships abroad - Research
• Start date: May 30, 2014
• End date: July 31, 2014
• Field of knowledge: Engineering - Biomedical Engineering - Bioengineering
• Principal Investigator: Maria Claudia Ferrari de Castro
• Grantee: Maria Claudia Ferrari de Castro
• Host Investigator: Dinesh Kant Kumar
• Host Institution: Centro Universitário FEI (UNIFEI). Campus de São Bernardo do Campo. São Bernardo do Campo , SP, Brazil
• Institution abroad: RMIT University, Melbourne, Australia

Abstract

Based on the fact that the command protocol used in the current upper limb myoelectric prostheses is not intuitive and does not mimic the biological movement control way, a lot of effort have been made in that field to correlate the control signal to the performed movement. Adding the fact that in situations where myoelectric signal has very low amplitude such as those observed in dynamic contractions to perform finger movements and hand gestures, without load, the techniques commonly presented in the literature do not reach good results. Thus, there is still need to investigate a robust source of information, suitable to representing the myoelectric signal in these conditions. Within this context, this work aims to continue the research line that has already begun, and proposes to investigate the techniques of the RMIT group involving the application of fractal dimension as technique for feature extraction process. Using common classifiers and the lowest possible number of channels, both the movement of individual fingers as a given set of hand gestures, essential in performing daily activities, can be distinguished. Some common gestures are an open hand, grasp a tripod pinch. (AU)