Application of chaotic time series analysis to signal characterization

Methods of chaotic time series analysis are applied to the problem of modeling nonlinear physical processes that occur in an emitter. Several approaches to the inverse problem in chaos theory are applied to pulse data having Gaussian noise and embedded nonlinear dynamics (Duffing´s equation) of appropriate scale. The nonlinear models created from the time series are able to separate the nonlinear dynamics characteristic of an emitter from the noise as long as the dimension of the nonlinear system is low. The numerical techniques used consist of the local prediction method of Farmer and Sidorowich and a neural network with radial basis functions. A comparison of the methods in terms of predictor errors, number of data points, and dimension of the Duffing attractor under various parameters is discussed