One of the methods of artificial hand
prosthesis control is the use of a surface electromyogram
signal. There are many methods to control the prosthesis,
each of which has advantages and disadvantages. In this
study, first, the surface electromyogram signal of people's
hands, which is related to the 6 movements that are the
most active in daily movements, is recorded and then
stored. To identify the movement pattern, 8 temporal
features are extracted from the signal, and then the best
feature is selected using a blind search of sources and
given to the input of the neural network. The results
showed that the average classification accuracy by
multilayer perceptron and PCA output is 96.77%, while
the average classification accuracy using all features is
82.77%.
Keywords :
Pattern recognition, hand movement detection, electromyogram signal.