Synchronization based system identification of an extended excitable system

Abstract

A basic state and parameter estimation scheme for an extended excitable system is presented, where time series from a spatial grid of sampling points are used to drive and synchronize corresponding model equations. Model parameters are estimated by minimizing the synchronization error. This estimation scheme is demonstrated using data from generic models of excitable media exhibiting spiral wave dynamics and chaotic spiral break-up that are implemented on a graphics processing unit.