Robust Control of a Prosthetic Hand Based on a Hybrid Adaptive Finger Angle Estimation

This paper presents a robust control approach to control the movement of a prosthetic hand based on an estimation of the finger angles using surface electromyographic (sEMG) signals. All the available prosthesis uses the motion control strategy which is pre-programmed get initiated when some threshold value of the measured sEMG signal is reached for a particular motion set. Here we use a novel approach to model the finger angle which utilizes System Identification (SI) techniques. The dynamic model obtained allows the instantaneous control for the finger motions. sEMG data is acquired using an array of nine sensors and the corresponding finger angle is acquired using a finger angle measuring device and a data glove. A nonlinear Teager–Kaiser Energy (TKE) operator based nonlinear spatial filter is used to filter sEMG data whereas the angle data is filtered using a Chebyshev type-II filters. An EMG-angle estimation model is proposed then the estimated angles are used to control to control movement of a prosthetic hand using a robust approach which can deal with modeling uncertainty. The overall performance of the prosthetic hand are measured based on numerical simulation. The resulting fusion based output of this approach plus the robust controller gives improved the prosthetic hand motion control. Key-Words: Robotics, Control, Prosthetics, Robust Control, sEMG, Finger Angle Model, System Identification.