EMG Based Classification of Hand Motions for Use in Robotics and Prosthetics
Location
Atrium
Session Format
Poster Presentation
Research Area Topic:
Engineering and Material Sciences - Electrical
Co-Presenters and Faculty Mentors or Advisors
Faculty Adviser: Dr. Rocio Alba-Flores
Abstract
Electromyography (EMG) is the study of electrical signals produced by the movement of muscles in the human body. EMG signals are used in many clinical and biomedical applications, such as in the diagnosis of neuromuscular diseases and rehabilitation through the control of assistive devices. However, surface electromyography (sEMG) presents a challenge to researchers because of their non-linear characteristics and ability to be easily effected by noise. In addition, sEMG from different subjects are also different from subject to subject because the signals are controlled by the nervous system and dependent on the anatomical and physiological properties of the muscles. However, sEMG are advantageous because they are non-invasive and relatively easy to obtain from subjects. Traditional processing of sEMG have included time-domain parameters such as mean absolute value, variance, zero-crossing, and average rectified value. Applications include myoelectric prostheses, robotics control, and virtual reality gaming controllers all of which utilize properly classified sEMG as control signals. Accurate and reliable classification of these signals has presented a challenge to the research community and prevents sEMG from being utilized as control signals in applications. This research presents the development of artificial neural networks (ANN) as a pattern recognition system to classify surface electromyography signals (sEMG) into hand motions. The main purpose of this research is to determine patterns and associations between the muscles in the forearm and related movements of the hand and fingers. When a pattern is determined, a database of the related signals will be developed for further research and development. The methods for this project include the use of an EMG machine that is already available in the EE department. EMG signals from different muscles in the forearm will be recorded while subjects are instructed to perform basic movements of individual fingers. These methods could then benefit society by applied to practical situations in the biomedical field such as developing adaptive prosthesis for different patients without the need of patient-specific training.
Keywords
Surface electromyography (sEMG), Artificial neural network, EMG classification, EMG database
Presentation Type and Release Option
Presentation (Open Access)
Start Date
4-24-2015 10:45 AM
End Date
4-24-2015 12:00 PM
Recommended Citation
Hickman, Stephen D., "EMG Based Classification of Hand Motions for Use in Robotics and Prosthetics" (2015). GS4 Georgia Southern Student Scholars Symposium. 32.
https://digitalcommons.georgiasouthern.edu/research_symposium/2015/2015/32
EMG Based Classification of Hand Motions for Use in Robotics and Prosthetics
Atrium
Electromyography (EMG) is the study of electrical signals produced by the movement of muscles in the human body. EMG signals are used in many clinical and biomedical applications, such as in the diagnosis of neuromuscular diseases and rehabilitation through the control of assistive devices. However, surface electromyography (sEMG) presents a challenge to researchers because of their non-linear characteristics and ability to be easily effected by noise. In addition, sEMG from different subjects are also different from subject to subject because the signals are controlled by the nervous system and dependent on the anatomical and physiological properties of the muscles. However, sEMG are advantageous because they are non-invasive and relatively easy to obtain from subjects. Traditional processing of sEMG have included time-domain parameters such as mean absolute value, variance, zero-crossing, and average rectified value. Applications include myoelectric prostheses, robotics control, and virtual reality gaming controllers all of which utilize properly classified sEMG as control signals. Accurate and reliable classification of these signals has presented a challenge to the research community and prevents sEMG from being utilized as control signals in applications. This research presents the development of artificial neural networks (ANN) as a pattern recognition system to classify surface electromyography signals (sEMG) into hand motions. The main purpose of this research is to determine patterns and associations between the muscles in the forearm and related movements of the hand and fingers. When a pattern is determined, a database of the related signals will be developed for further research and development. The methods for this project include the use of an EMG machine that is already available in the EE department. EMG signals from different muscles in the forearm will be recorded while subjects are instructed to perform basic movements of individual fingers. These methods could then benefit society by applied to practical situations in the biomedical field such as developing adaptive prosthesis for different patients without the need of patient-specific training.