Estimating model parameters from time series by autosynchronization.

s The synchronization of (unidirectionally) coupled dy namical systems and its possible applications in commu cation schemes is currently a field of great interest ( [1–7] and references cited therein). In this Letter w discuss a special feature of synchronizing systems ca autosynchronizationwhere a system with slowly varying parameters converges from a state of nonsynchroniza to synchronization. This adaption process is governed additional ordinary differential equations (ODEs) for th parameters that are controlled by the synchronization er A systematic way for deriving the parameter controllin loop is presented and illustrated by numerical examp For the sake of brevity we consider unidirectionally co pled systems only, although the main ideas can in princ also be applied to mutually coupled synchronizing syste In order to indicate a possible application in nonlinear tim series analysis [8] and system identification autosynch nization is discussed and used in the following for es mating the parameters of a given model from a scalar ti series [9–14]. Let