True random number generator design based on the fractional-order Sprott H chaotic system with statistical validation

Abstract

The increasing demand for secure communication systems has emphasized the necessity of high-quality entropy sources in cryptographic applications. True Random Number Generators (TRNGs), which derive randomness from physical and chaotic processes, are essential for ensuring data confidentiality in domains such as the Internet of Things (IoT), healthcare, and wireless communication. This study presents a novel TRNG architecture based on the Fractional-Order Sprott H Chaotic System (FOSHCS), a model not previously employed in TRNG design. The chaotic properties of FOSHCS were rigorously evaluated through bifurcation diagrams, the maximum Lyapunov exponent (MLE), and attractor projections, confirming its viability as a reliable entropy source. The system was physically implemented on an NVIDIA Jetson AGX Orin platform using a custom-designed DAC circuit to observe the chaotic trajectories in the analog domain. Furthermore, real-time GPU temperature data was incorporated with the chaotic output to enhance entropy diversity. The resulting bitstreams underwent standard statistical randomness tests, including the NIST SP 800-22, FIPS 140-1, and ENT test suites, all of which were successfully passed. The integration of fractional-order chaotic modeling with physical entropy harvesting enabled the development of a compact and high-entropy TRNG suitable for embedded and security-critical applications. To the best of our knowledge, this work represents the first hardware realization of a TRNG based on the FOSHCS, offering a promising new direction in secure and robust random number generation.