Graphical simulation of artificial hand motion with fuzzy EMG pattern recognition

A graphical simulator in which a prosthesis is controlled by means of electromyogram (EMG) processing is described. The integral of the absolute value (IAV) of the biceps and triceps EMGs are used as features. A fuzzy k-means scheme is used to classify the motion before actuating a 3-degrees-of-freedom arm graphically on a computer monitor. The main advantages of such an approach over a classical training are: (i) the possibility of training amputees before using a real prosthesis, (ii) its cost-effectiveness by manpower time-saving, (iii) greater availability of the training set, and (iv) self-paced learning.