V.V. Afanasiev, S.S. Loginov

Development of new data carriers on the basis of non-linear system characterised by dynamic chaos is also one of the most important practical targets. The overall concern about the chaotical signals is first of all determined by their correlating and spectral characteristics similar to the characteristics of random signals. Analogous realizations forming pseudorandom signals on the basis of non-linear systems with chaotical dynamics do not possess any reproduced statistical characteristics. Thus it makes it necessary to form pseudorandom signals on the basis of numerical integration procedures of non-linear difference equations used to describe systems with chaotic dynamics. The aim of the work is to develop of new data carriers on the basis of non-linear system by researching probability characteristics depending on discretisation step size. Temporal discretisation step estimation ratio to quasi-resonant period oscillations was varied from 100 to 1000. Probability distributions of the Lorenz system signals are considered. Discretization step size influence on poly Gauss properties of signals was carried out. Minimum number of samples giving robustness of poly Gauss statistics was got. Measure of inaccuracy of approximation no more than 5% for X,Y signals of Lorenz system and no more than 7% for Z signals of Lorenz system. Robust distributions period formation is depend on temporal discretisation step estimation ratio to quasi-resonant period oscillations. Estimations of the minimum number of points where robust distributions received was got.