A High-Performance FPGA PRNG Based on Multiple Deep-Dynamic Transformations.

Area of Science:

• Information Security
• Applied Mathematics
• Computer Engineering

Background:

• Pseudo-random number generators (PRNGs) are crucial for cryptography, statistical analysis, and emerging fields like AI and big data.
• The increasing demand for high-speed, high-quality random number generation necessitates innovative solutions.

Purpose of the Study:

• To develop a novel PRNG algorithm addressing the growing need for high-speed, high-quality random number generation.
• To implement and evaluate the performance of the proposed PRNG for information security applications.

Main Methods:

• Developed a multiple deep-dynamic transformation (MDDT) algorithm integrated with a skewed tent map for complex dynamics.
• Implemented the improved one-dimensional discrete chaotic mapping on a Xilinx xc7k325tffg900-2 Field-Programmable Gate Array (FPGA).
• Validated the PRNG's quality using NIST SP800-22, diehard, and TestU01 test suites.

Main Results:

• The proposed PRNG successfully passed all NIST SP800-22, diehard, and TestU01 statistical tests.
• Achieved high novelty performance operating at a 150 MHz clock frequency.
• Reached a maximum throughput of 14.4 Gbps, surpassing most existing studies.

Conclusions:

• The MDDT-based chaotic PRNG offers superior performance and high novelty.
• The FPGA implementation is efficient and suitable for demanding embedded applications in information security.
• This work contributes a robust solution for high-speed pseudo-random number generation.