Abstract

The article deals with the formation of phase portraits patterns of the correlation function of uncertain time series. Those patterns are used in expert software systems for determination of hidden regularities in random processes. Problem of the identification and classification of the irregular random events is particularly relevant in the case of complete absence of a priori information about the investigated process. This problem becomes much more complicated in case of inability to repeat the experiment. The methods of nonlinear dynamics for the hidden patterns identification, which may significantly affect the behavior of the system and the change nature of the phase trajectories, are proposed. A correlation function is proposed as a behavioral characteristic of informative systems because it fully displays properties of time series. The phase portrait of correlation function is used as an indicator of hidden events. The structure of simulation modeling software bundle, which implements the selected algorithm of phase portrait building, is described. The implementation of some modules in LabView, which allows the creation of different types of phase portraits for specific application, is provided. The database of reference signal patterns, their spectral and correlation characteristics is created. On this basis, a set of phase portraits templates was developed. Particular attention is paid to the model of the signal providing the turbulence regime. This paper contains partial results of model experiments that confirmed the possibility of using this approach for the phase portraits modeling which are similar to the phase portraits of real time series. Results obtained from studies can be used practically for analysis of undetermined data from dynamical systems analysis, which model is presented as a “black box” with single output.

Authors and Affiliations