Аналіз та реалізація справжніх генераторів випадкових чисел на основі архітектури FPGA

Abstract

True random number generators (TRNGs) are essential for cryptographic systems, secure communication protocols, and stochastic modeling. Unlike pseudo-random generators, TRNGs obtain randomness from physical processes, ensuring higher unpredictability. This paper presents the analysis and implementation of a TRNG based on field-programmable gate array (FPGA) architecture. The proposed design uses ring oscillators as entropy sources, exploiting phase jitter caused by intrinsic noise in silicon structures. To increase entropy quality, outputs from multiple oscillators are combined using an XOR-based postprocessing stage. Special synthesis constraints are applied to preserve combinational loops during hardware implementation. The generated bitstreams are evaluated using statistical tests from the NIST SP 800-22 suite. The results confirm that FPGA-based TRNGs can produce high-quality random numbers suitable for cryptographic applications