Security analysis of a random number generator based on a chaotic hyperjerk system

Abstract

This paper analyzes the security of a random number generator (RNG) based on a 4-D chaotic hyperjerk system. An attack system is designed to reveal the security weaknesses of the proposed chaotic RNG. Knowing the structure of the RNG and observing one of the state variables of the chaotic system, convergence between attack and target systems is demonstrated by applying linear continuous coupling in master-slave synchronization scheme. Output bit sequence of the chaotic RNG is identically reproduced. The feasibility of the attack system is verified through numerical simulations. In this paper, a specific continuous-time chaos-based RNG is targeted as a case study. However, the cryptanalysis method presented in this paper is applicable to any continuous-time or discrete-time chaos-based RNGs. Therefore, this study highlights the security vulnerabilities of chaos-based RNGs and underlines that deterministic chaos itself cannot be considered as an entropy source for generation of random numbers.