Abstract

A human wearing an exoskeleton-type assistive device results in a parallel control system that includes two controllers: the human brain and a digital exoskeleton controller. Unknown and complicated characteristics of the brain dynamically interact with the exoskeleton controller which makes the controller design challenging. In this paper, the motion control system of a human is regarded as a feedback control loop that consists of a brain, muscles and the dynamics of the extended human body. The brain is modeled as a control algorithm amplified by a fictitious variable gain. The variable gain compensates for characteristic changes in the muscle and dynamics. If a human is physically impaired or subjected to demanding work, the exoskeleton should generate proper assistive forces, which is equivalent to increasing the variable gain. In this paper, a control algorithm that realizes the fictitious variable gain is designed and its performance and robustness are discussed for single-input single-output cases. The control algorithm is then verified by simulation results.