High-Entropy True Random Number Generator Based on Memristor Reset Switching

Aside from storing data, memristors can also be used as an entropy source, typically by utilizing the random characteristics of the high resistance state (HRS) or set/reset time delay (TD) of memristors. However, obtaining a reliable high-entropy source remains difficult. In this letter, we designed a novel random number generator (TRNG) circuit that takes full advantage of the memristor's randomness of reset switching, combining the HRS and TD entropies to achieve higher entropy. The entire random characteristic is derived from two temporal processes: the charging and discharging of a capacitor connected to the memristor. Experiments and simulations illustrate the function and advantages of the TRNG circuit, and a min-entropy of 0.9989 is attained, which is higher and more robust than two independent sources, and which can be further enhanced by the device design itself.