Design and Development of a Finger Exoskeleton for Motor Rehabilitation using Electromyography Signals

Abstract

Robotic devices can accelerate the rehabilitation process of an impaired patient by increasing the frequency of treatment and eliminating the need for a physiotherapist. This paper presents the development of a novel human finger exoskeleton for the motor rehabilitation of post-stroke patients by providing assistance-as-required support to patients in active flexion of the finger. The proposed design comprises of a two degrees of freedom, seven-link concatenated mechanism with two links attached to proximal and distal phalanges and actuated by two electric linear actuators for compact size. We synthesized the mechanism on an interactive geometry software to make the mechanism follow natural joint trajectories. Kinematic analysis of the exoskeleton showed very low deviation from ideal trajectory and finite element analyses showed good structural rigidity under maximum load conditions. The paper further describes the electro-mechanical system and software architecture of the prototype. The prototype makes use of electromyography signal to detect human intention and has the ability to provide assistance-as-required rehabilitation through a developed intention based control algorithm. The algorithm makes use of signal processing and statistical techniques to recognise patterns from electromyography data. A classifier was modelled with human intention detection accuracy of nearly 90% The results of the testing on real human finger demonstrate the potential for clinical applications in the robot-assisted rehabilitation of the human hand for grasping tasks.