Researchers in Brazil and Australia have established that to reliably control a myoelectric prosthetic hand, the gestures communicated from the forearm muscles to the hand should be carefully selected to produce the most accurate recognition. The study also developed a method to select the most reliable finger and hand positioning gestures and that using the selection process “lowers the misclassifications and improves the accuracy leading to greater reliability and safety for the user.”
Experiments recorded the surface electromyogram (sEMG) signals from the forearms of four participants as they performed five finger positioning gestures and five functional hand grips. The movements were repeated thirty times in random order. The results were then classified off-line, and the gestures were ranked using a proposed positive-negative performance measurement (PNM) index, according to the authors.
When using ten gestures, the sensitivity rate was 80 percent and the specificity rate was 97.8 percent. After ranking the gestures using the PNM, a subset of six gestures-hand open, hand close, little finger flexion, ring finger flexion, middle finger flexion, and thumb flexion-provided sensitivity and specificity rates of 96.5 percent and 99.3 percent.
The study’s authors note that further testing is required to generalize the principle, which may result in a different number of gestures that provide the best sensitivity and specificity.
The open-access article was published April 9 in the journal BioMedical Engineering OnLine.
