Proportional and Simultaneous Control Strategy for 3 DoF of Hand

Abstract

Amputation is a physical deficiency affecting indiscriminately the world population, representing a high cost for social security systems in health, in order to improve the life quality of amputees, is necessary obtain prosthesis that simulate the movement of the hand as naturally as possible. This paper, shows a methodology for the continuous and simultaneous control of 3 degrees of freedom (DoF) of the hand, for further applications on prostheses control. The position of the flexion/extension of the wrist (DoF1), Metacarpophalangeal (MCP) joint for tri-digital grasp (DoF2) and penta-digital grasp (DoF3) were estimated, based on surface Electromyography (sEMG) and Support Vector Regression (SVR), using time domain features widely applied in the processing of sEMG signals, like the Hudgin’s feature set. Intersubject and Intrasubject model were assessed, finding that Intrasubjects models had the best performance with R $$^2$$ of 0.73, 0.49 and 0.51 for DoF1, DoF2 and DoF3, respectively. Our approach was able to estimate the target motion profile within $$\mathbf {~22.5^{\circ }}$$ with a low memory usage and minimum computational complexity.