Function Composition from Sine Function and Skew Tent Map and Its Application to Pseudorandom Number Generators

• Main Authors: Leonardo Palacios-Luengas, Ricardo Marcelín-Jiménez, Enrique Rodriguez-Colina, Michael Pascoe-Chalke, Omar Jiménez-Ramírez, Rubén Vázquez-Medina
• Format: Article
• Language: English
• Published: MDPI AG 2021-06-01
• Series: Applied Sciences
• Subjects: pseudorandom number generator; function composition; nonlinear dynamics and chaos; robust chaotic map
• Online Access: https://www.mdpi.com/2076-3417/11/13/5769

In cryptography, the pseudorandom number sequences must have random appearance to be used in secure information systems. The skew tent map (STM) is an attractive map to produce pseudorandom sequences due to its easy implementation and the absence of stability islands when it is in chaotic behavior. Using the STM and sine function, we propose and analyze a function composition to propose a pseudorandom number generator (PRNG). In the analysis of the function composition, we use the bifurcation diagram and the Lyapunov exponent to perform a behavioral comparison against the STM. We show that the proposed function composition is more sensitive to initial conditions than the STM, and then it is a better option than the STM for cryptography applications. For the proposed function we determine and avoid the chaos annulling traps. The proposed PRNG can be configured to generate pseudorandom numbers of 8, 16 or 32 bits and it can be implemented on microcontrollers with different architectures. We evaluate the pseudorandomness of the proposed PRNG using the NIST SP 800-22 and TestU01 suites. Additionally, to evaluate its quality, we apply tests such as correlation coefficient, key sensitivity, statistical and entropy analysis, key space, linear complexity, and speed. Finally, we performed a comparison with similar PRNGs that produce pseudorandom sequences considering numbers of 8 and 32 bits. The results show that the proposed PRNG maintains its security regardless of the selected configuration. The proposed PRNG has five important features: easy implementation, configurable to produce number with 8, 16 or 32 bits, high processing speed, high linear complexity, and wide key space. These features are necessary for cryptographic systems.