Simulation of chaotic-based random number generators in encryption applications using echo state networks

Random number generators (RNGs) are fundamental components of encryption algorithms, playing a crucial role in secure communication and data protection processes. Traditional random number generation methods may fail to meet certain statistical and security criteria, whereas chaotic systems, with their deterministic yet unpredictable dynamics, offer a promising alternative with strong randomness properties. In this thesis, the reliability and performance of chaotic-based random number generators (CRNGs) are examined, and their simulation is carried out using reservoir computing (RC). Reservoir computing, with its low computational cost and powerful nonlinear dynamics, serves as an effective tool for modeling chaotic systems. In this study, the statistical properties of random numbers generated by different chaotic maps and systems are analyzed, and their accuracy is evaluated using reservoir computing techniques. The results demonstrate that chaotic-based random number generators can be effectively used in high-security encryption applications. Moreover, the study reveals that reservoir computing can be successfully applied to mimic chaotic systems and analyze randomness. By exploring the relationship between chaotic systems and artificial neural networks, this research contributes to the development of more secure encryption systems.