Two-dimensional myoelectric control of a robotic arm for upper limb amputees

Abstract

Rehabilitation engineering and medicine have become integral and significant parts of health care services, particularly and unfortunately in the last three or four decades, because of wars, terrorism and large number of car accidents. Amputees show a high rate of rejection to wear prosthetic devices, often because of lack of an adequate period of adaptation. A robotic arm may appear as a good preliminary stage. To test the hypothesis, myoelectric signals from two upper limb amputees and from four normal volunteers were fed, via adequate electronic conditioning and using MATLAB, to an industrial robotic arm. Proportional strength control was used for two degrees of freedom (x-y plane) by means of eight signal features of control (four traditional statistics plus energy, integral of the absolute value, Willison's amplitude, waveform length and envelope) for comparison purposes, and selecting the best of them as final reference. Patients easily accepted the system and learned in short time how to operate it. Results were encouraging so that valuable training, before prosthesis is implanted, appears as good feedback; besides, these patients can be hired as specialized operators in semi-automatized industry.