Performance of a myoelectric controller for an elbow-wrist exoskeleton for incomplete spinal cord injury subjects

摘要

Electromyographic (EMG) control interfaces have the potential to increase the effectiveness and accessibility of rehabilitation robotics to a larger population of impaired individuals, including those with no residual motion in the upper limb. In a previous study, the authors characterized surface EMG patterns for able-bodied and incomplete spinal cord injury (iSCI) subjects while wearing an upper-limb exoskeleton and verified that a pattern recognition approach to detecting the intent of iSCI users is viable. EMG features were selected as inputs to a classifier of the user's intended direction of movement. Training and testing of classification was done offline using linear discriminant analysis (LDA). The goal of the current experiment is to extend the use of this controller to a real-time setting.