Design of a Cable-Driven Arm Exoskeleton (CAREX) for Neural Rehabilitation

Ying Mao, Sunil Kumar Agrawal

    Research output: Contribution to journalArticle

    197 Citations (Scopus)

    Abstract

    Rehabilitation robots are, currently, being exploredfor training of neural impaired subjects or for assistance of thosewith weak limbs. Intensive training of neurally impaired subjects,with quantifiable outcomes, is the eventual goal of these robot exoskeletons.Conventional arm exoskeletons for rehabilitation arebulky and heavy. In recent years, the authors have proposed tomake lightweight exoskeletons for rehabilitation by replacing therigid links of the exoskeletonwith lightweight cuffs fixed to themovinglimb segments of the human arm. Cables are routed throughthese cuffs, which are driven by motors, to move the limb segmentsrelative to each other. However, a scientific limitation of acable-driven system is that each cable can only pull but not push.This paper is the first to demonstrate via experiments with cabledrivenarm exoskeleton (CAREX) that it is possible to achieve desiredforces on the hand, i.e., both pull and push, in any direction as requiredin neural training. In this research, an anthropomorphicarm was used to bench test the design and control concepts proposedin CAREX. As described in this paper, CAREX was attachedto the limb segments of a five degree-of-freedom anthropomorphicarm instrumented with joint sensors. The cuffs of CAREX weredesigned to have adjustable cable routing points to optimize the“tensioned” workspace of the anthropomorphic arm. Simulationresults of force field for training and rehabilitation of the arm arefirst presented. Experiments are conducted to show the performanceof a CAREX force field controller when human subjectspull the end-effector of the anthropomorphic arm to travel on prescribedpaths. The human–exoskeleton interface is also presentedat the end of this paper to demonstrate the feasibility of CAREXon human arm.
    Original languageEnglish
    Pages (from-to)922-931
    JournalIEEE Transactions on Robotics
    Volume28
    Issue number4
    DOIs
    Publication statusPublished - Aug 2012

      Fingerprint

    Cite this