Non-linear modeling of human muscle in functional electrical stimulation with machine learning

Abstract

There are severe paralyzed people by accidents who have difficulty to move paralyzed parts of bodies. Function electrical stimulation (FES) assists motion of body because electrical stimulation contracts muscles. FES has been studied for a long time. However, an accurate dynamic model of body does not exist. In this paper, the dynamic model is given by a neural network trained the relationship between human's joint acceralation and ampulitude voltage.