METHODS FOR IMPROVING THE CRYPTOGRAPHIC STRENGTH OF PSEUDORANDOM SEQUENCES BASED ON CHAOTIC GENERATORS

Abstract

The paper investigates methods for improving the cryptographic strength of pseudorandom bit sequences generated using software-based chaotic generators. The possibility of employing chaotic oscillations as a source of deterministic unpredictability is considered, combining high sensitivity to initial conditions, a wide spectral range, and statistical properties close to those of random processes. This approach makes it possible to generate sequences with increased entropy, which is a key requirement for cryptographic applications. Based on a comparative analysis of numerical and binary sequences formed using the logistic map in the chaotic regime, the influence of quantization methods and additional processing on the statistical and correlation characteristics of pseudorandom sequences is studied. It is shown that the use of the sample mean as a quantization threshold ensures a balance between zeros and ones even in the presence of statistical deviations typical for practical implementations of chaotic generators. To further increase unpredictability and improve the spectral properties of the generated sequences, a cyclic shift operation is applied. The relevance of this research is justified by the expediency of using chaotic generators in information security systems operating under conditions of active electronic countermeasures and intentional interference with communication channels. Under such conditions, classical cryptographic algorithms require enhancement by methods capable of providing both cryptographic strength and concealment of signal structures. The main objectives of the study include the generation of pseudorandom bit sequences based on chaotic mappings, evaluation of their statistical properties, and analysis of autocorrelation characteristics in order to determine their suitability for cryptographic and telecommunication applications. Autocorrelation analysis confirmed a low level of correlation dependencies and the absence of pronounced periodicities in the generated sequences.

The obtained results can be used in the development of cryptographic systems, spread spectrum technologies, as well as signal-code constructions for secure communication channels, particularly under conditions of electronic warfare.