Active tracking movements of flexion and extension on the elbow joint for an initial prototype of EMG-driven assistant exoskeleton

Abstract

Various exoskeletons have been developed for the assistance of upper and lower limbs rehabilitation, however, therapists and doctors have pointed out that passive-repeating motions of limbs is ineffective in terms of neurological recovery. To improve the neurological recovery of human motor control, active-assisted encouragement as therapist and doctors manually provided should also be simulated with help of exoskeleton targeting for rehabilitative purpose. In this paper, a single-degree-of-freedom (SDOF) robotic joint to mimic elbow movements is developed and EMG signals are collected using as a bio-feedback signal. A pair of muscles including biceps and triceps that involved in flexion and extension on the elbow joint are selected, acquisition and processing of EMG signals is then achieved by designed hardware filtering and proposed software algorithms. By analysis of the resultant EMG signals, the intentions of flexion and extension movements are quantitatively presented. Online control commands for the exoskeleton are then spontaneously produced. A close-loop control with EMG-driven in real time is applied on an initial prototype of human-machine interface (HMI) for an active assistance using the exoskeleton. Initial applications on the experimental prototype have proved the effectiveness of the proposed active tracking method that has successfully tracking effort that bridge the wearer and the wearable exoskeleton together.