Reconstruction of nonlinear time-delayed feedback models from optical data

We propose a new technique for the reconstruction of nonlinear time-delayed feedback models from a time series. Such delayed feedback behavior arises frequently in nonlinear optics. By the use of optimal transformations and maximal correlation, the delay time and the governing differential equation can be estimated with high accuracy. We demonstrate the applicability to chaotic and noisy data by the analysis of a passive optical resonator system, the Ikeda model. (C) 1999 Elsevier Science Ltd. All rights reserved.