Pattern Recognition of Individual and Combined Fingers Movements Based Prosthesis Control Using Surface EMG Signals

Prosthesis control system is the need for the amputees or disable person for performing their daily household work and interaction with the outside world. It is the fundamental component of modern prostheses, which uses the myoelectric signals from an individual’s muscles to control the prosthesis movements. The surface electromyogram signals (SEMG) being noninvasive has been used as a control source for multifunction powered prostheses controllers. In spite of the fact there is wide research on the myoelectric control of movements of forearm and hand movements but a little research has been carried out for control of more dexterous individual and combined fingers. With the current demands of such prostheses a challenge that exists is the ability to precisely control a large number of individual and combined finger movements and that too in a computationally efficient manner. This paper investigates accurate and correct discrimination between individual and combined fingers movements using surface myoelectric signals, in order to control the different finger postures of a prosthetic hand. We have SEMG datasets with eight electrodes located on the human forearm and fifteen classes. Various feature sets are extracted and projected in a manner to ensure that maximum separation exists between the finger movements and then fed to the four different classifiers. Practical results along with the statistical significance tests proved the feasibility of the proposed approach with mean classification accuracy greater than 95% in finger movement classification.