随机数发生器的噪声模型与快速仿真

在随机数发生器设计过程中,采用Spice虽然能对模拟电路进行比较精确的仿真,但耗时太久,同时它的数据处理工作也非常繁琐,不易进行统计分析。为了解决这一问题,在分析了一个分段线性映射的、决定论混沌的随机数发生器的基础上,根据它的混沌公式提出了基本的仿真原理。为了提高仿真的准确性,针对CMOS电路中主要的噪声,即MOS管的热噪声与闪烁噪声,建立了二种噪声仿真模型。在分析电路行为的同时,逐步提出了采用MATLAB/Simulink的随机数发生器单元电路模型,然后建立了随机数发生器系统模型,实现了快速计算过程。最后给出了系统模型的仿真测试结果。

The Noise Model and Fast Simulation of A Random Number Generator

In the design of Random Number Generator (RNG), Spice simulation of the whole analog circuit is fairly accurate but costs a lot of time; also the generated data are hard for post-processing and analyzing. To solve the problem, the paper analyzes a deterministic chaos RNG based on Piecewise Linear Map, and then presents the principle for simulation according to its chaotic equation. To improve the accuracy, two noise models (thermal and flicker, which are the dominating noise sources in CMOS circuits) have also been presented. After analyzing the behavior of the RNG circuit, a MATLAB/Simulink model for RNG cell is provided, and a system model of the RNG is built thereafter. Those models have made the fast simulation realizable. Testing results of the RNG system are given at the end.