基于混沌的高速随机数发生器

A High Speed Chaos-Based Random Number Generator

基于混沌的随机数发生器采用了离散时间的决定论混沌系统。决定论混沌的一个本质特征是对初始值的敏感依赖性。由于初始值是一个模拟电路的初值,对于数字测量系统是永远无法逼近或达到的,它的偏差使得测量系统产生的符号序列以后有着充分大的分离,从而使得符号序列不可预知、不可再现,具有真随机的特性。在分析了一类分段线性映射的决定论混沌系统的基本特性后,设计了由开关电容电路等组成的模拟电路。为了保证随机序列的分布特性,针对CMOS电路中主要的噪声,即MOS管的热噪声与闪烁噪声,设计时建立了二种噪声仿真模型;同时为了加快分析的效率和速度,提出了一种快速分析方法。最后,采用NIST标准进行了测试。

A deterministic chaotic discrete time dynamical system can be used for random number generator (RNG). Initial condition sensitivity is the key characteristic of deterministic chaos. As initial condition is an analog value in the circuit, it will never be approached or reached in the digital measuring system. Small difference of initial value between symbol sequences generated by the measuring system will develop into great difference between later items of symbol sequences as to lead two sequences separate fully. Therefore, the symbol trajectory will never be predicted or reproduced. After analyzing the basic characteristic of a Piecewise Linearity Mapping, a design of analog circuits including switched capacitors is presented. To verify the expected distribution of the random number generated by the circuit, the dominating CMOS noise was studied with two noise models (thermal and flicker). To speed up the circuit simulation, a fast simulation method is proposed. Finally, the experimental result of the RNG in accordance with NIST (National Institute of Standards and Technology) standard was provided.