Title: The identification of continuous, spatiotemporal systems

Abstract: We present a method for the identification of continuous, spatiotemporal dynamics from experimental data. We use a model in the form of a partial differential equation and formulate an optimization problem for its estimation from data. The solution is found as a multivariate nonlinear regression problem using the ACE-algorithm. The procedure is successfully applied to data, obtained by simulation of the Swift-Hohenberg equation. There are no restrictions on the dimensionality of the investigated system, allowing for the analysis of high-dimensional chaotic as well as transient dynamics. The demands on the experimental data are discussed as well as the sensitivity of the method towards noise.