Skip to Main Content
The myoelectric signal has played a major role in the development of prosthesis control technology. A myoelectric classification system has the ability to determine a prosthesis user's intent based solely on his or her muscle activity, thereby allowing for more intuitive prosthetic control. Much work has been done on the recognition of upper arm and gross hand movement tasks, but it was not until accuracy levels approached 100%  that more attention was given to specific finger movements. In this study, the effect of electrode array size and arrangement on classification accuracy is investigated for a four-finger typing task. This follows from previous work  in which the classification system itself was optimized. Unique advantages were found using array sizes of three and seven electrodes; classification accuracy of 92.7plusmn3.9% was found in the latter case across twelve subjects.
Date of Conference: 3-6 Sept. 2009