NARX Recurrent Neural Network Model of the Graphene-Based Electronic Skin Sensors with Hysteretic Behaviour

Abstract

AbstractThe electronic skin described in the article comprises screen-printed graphene-based sensors, intended to be used for robotic applications. The precise mathematical model allowing the touch pressure estimation is required during its calibration. The article describes the recurrent neural network model for graphene-based electronic skin calibration, in which parameters are not homogeneous, and the touch force characteristics have visible hysteretic behaviour. The presented method provides a simple alternative to the models known in the literature.