Controle de força de uma prótese mioelétrica de mão com realimentação sensorial

Abstract

Hand myoelectric prostheses that use sensory feedback can provide improvement in fragile objects manipulation and in user’s acceptance. The force sensory feedback can be done in many ways. The goal of this study was to develop a force control system for a myoelectric hand prosthesis with visual feedback and by sinusoidal electric stimulation of tactile fibers. Surface Electromyographic (sEMG) signals were recorded in healthy participants, performing sustained contractions of hand grasp, in different force levels. Those signals were used in the implementation of a force estimator considering two methods. A proportional control with derivative feedback was also implemented for the prosthesis. The interface with the electrical stimulator was developed for the force feedback. The force estimator with best performance was the one based in an artificial neural network for function approximation. The closed-loop control system used only the visual feedback, once the sinusoidal current feedback generated interference in the sEMG signals. The force control system was tested by 5 healthy participants and obtained Correlation Coefficient of 89,9% in the online force estimation with control’s success rate of 67,66%. Thus, both the force estimator as the control of the prosthesis have shown to be promising