Fractional-Order Modeling and Identification for a Phantom EEG System

Abstract

This paper is about dynamical modeling of an electroencephalographic (EEG) measurement chain. It is based on a home-made experimental setup consisting of an electrolytic medium, stimulated by external electrodes, the activity of which is measured at different points via other electrodes, connected to an EEG recording device. A novel noninteger-order model structure is then established with a so-called constant phase element, and a Nyquist-based graphical method is derived to identify the noninteger order. The study is completed with a recursive identification method for full parameter estimation in the model, and various applications on experimental data are presented, finally resulting in a fully identified noninteger order high-pass filter.