Finger Movement Classification Using Forearm EMG Signals
Loading...
Authors
Andrews, Alexander
Date
2008-11-04T14:51:10Z
Type
thesis
Language
eng
Keyword
EMG , myoelectric , classification , pattern recognition , prediction , finger , typing , prosthesis
Alternative Title
Abstract
To a person with an upper limb amputation or congenital defect, a well-functioning prosthesis can open the door to many work and life opportunities. A fundamental component of many modern prostheses is the myoelectric control system, which uses the myoelectric signals from an individual's muscles to control prosthetic movements. Though much research has been done involving the myoelectric control of arm and gross hand movements, more dexterous finger control has not received the same attention. Consequently, the goal of this study was to determine an optimal approach to the myoelectric signal classification of a set of typing motions. Two different movement sets involving the fingers of the right hand were tested: one involving digits two through five (4F - "four finger"), and the other involving digits one and two (FT - "finger/thumb").
Myoelectric data were collected from the forearm muscles of twelve normally-limbed subjects as they performed a set of typing tasks. These data were then used to test a series of classification systems, each comprising a different combination of system element choices. The best classification system over all subjects and the best classification system for each subject were determined for both movement sets.
The optimal subject-specific classification systems yielded classification accuracies of 92.8 ± 2.7% for the 4F movement set and 93.6 ± 6.1% for the FT movement set, whereas the optimal overall classification systems yielded significantly lower performance (p<0.05): 89.6 ± 3.4% for the 4F movement set and 89.8 ± 8.5% for the FT movement set. No significant difference in classification accuracy was found between movement sets (p=0.802). A two-way repeated measures ANOVA (α=0.05) was used to determine both significance results.
Description
Thesis (Master, Electrical & Computer Engineering) -- Queen's University, 2008-10-31 14:59:43.151
Citation
Publisher
License
This publication is made available by the authority of the copyright owner solely for the purpose of private study and research and may not be copied or reproduced except as permitted by the copyright laws without written authority from the copyright owner.