True random number generator using stochastic noise signal of memristor with variation tolerance

Dayeon Yu,Suhyeon Ahn,Sangwook Youn,Jinwoo Park,Hyungjin Kim
DOI: https://doi.org/10.1016/j.chaos.2024.115708
2024-11-12
Abstract:Memristors are suitable for internet of things (IoT) edge devices due to their high scalability and low power consumption. Also, their stochastic noise signals can be used to produce entropy sources for information encryption in edge devices, which is essential during data exchange. Random telegraph noise (RTN) in memristors is a current fluctuation that occurs when electrons are captured or emitted from a trap, and its stochastic characteristics can be used as an entropy source for a true random number generator (TRNG). However, post-processing is essential because RTN tends to be biased toward specific time constants, and there is significant variation within each time constant. In this work, we present a TRNG circuit that can tolerate the time variation of RTN and demonstrate experimental results with a breadboard. The dependence of RTN signals on a read voltage is statistically verified, confirming a significant variation in a time constant of RTN signal (σ/μ > 700 %). Despite this variation, high randomness of generated random number stream can be obtained by a falling edge detector and is verified by NIST SP 800–22 test (with >14 tests passed) and auto-correlation function test (rate of exceeding confidence bound <1.5 %).
mathematics, interdisciplinary applications,physics, mathematical, multidisciplinary
What problem does this paper attempt to address?