Digital-Analog Hybrid Scheme and Its Application to Chaotic Random Number Generators.
Zeshi Yuan,Hongtao Li,Yunchi Miao,Wen Hu,Xiaohua Zhu
DOI: https://doi.org/10.1142/s0218127417502108
IF: 2.45
2017-01-01
International Journal of Bifurcation and Chaos
Abstract:Practical random number generation (RNG) circuits are typically achieved with analog devices or digital approaches. Digital-based techniques, which use field programmable gate array (FPGA) and graphics processing units (GPU) etc. usually have better performances than analog methods as they are programmable, efficient and robust. However, digital realizations suffer from the effect of finite precision. Accordingly, the generated random numbers (RNs) are actually periodic instead of being real random. To tackle this limitation, in this paper we propose a novel digital-analog hybrid scheme that employs the digital unit as the main body, and minimum analog devices to generate physical RNs. Moreover, the possibility of realizing the proposed scheme with only one memory element is discussed. Without loss of generality, we use the capacitor and the memristor along with FPGA to construct the proposed hybrid system, and a chaotic true random number generator (TRNG) circuit is realized, producing physical RNs at a throughput of Gbit/s scale. These RNs successfully pass all the tests in the NIST SP800-22 package, confirming the significance of the scheme in practical applications. In addition, the use of this new scheme is not restricted to RNGs, and it also provides a strategy to solve the effect of finite precision in other digital systems.