True Random Number Generators Based on MRG with Chaotic Seed Generation (I)

Information security is very important nowadays, especially for the personal privacy and personal information security. Recently, the great need offered by high-speed wireless communications in hospital, including uploading of X-ray scan, Endoscopy, CT, and MRI image of patients has stimulated growing interest in nonlinear PRNGs for security-related applications. Therefore, we need the fast and secure key generation for meeting the great requirement. It is a challenge to develop nonlinear PRNGs with high throughput rates and good unpredictability. On the one hand, pseudo random number generators (PRNGs) have been widely used in statistic simulations, Monte Carlo simulations, modeling, noise generating and telecommunication systems. For Monte Carlo simulations, Linear Congrent Generators (LCGs) and Multiplr Recursive Generators(MRGs) play an important role in generating fast and good random numbers. However, both of LCGs and MRGs are not secure generator because the next value is computed from the previous value(s) via a linear recurrence. On the other hand, true random number generators (TRNGs) amplify the disturbance in real world, such as thermal noise, super conductivity, and jitter in phase locked loop (PLL) to preserve the unpredictability. Their limitations are the slow generation and highly sensitive to the environment. In other words, if the huge noise comes, it is easy in TRNG to have bias in its random number generation mechanism. In that case, the generated numbers are not “random” anymore. We want to utilize the unpredictability in jitters of phase locked loop (PLL) for seed generation. Then put these unpredictable seed in our MRGs for random number generation. Before the attackers collect enough generated numbers, we reseed our MRGs with the chaotic seed generation to reduce the prediction probability. In this proposal, we will work on the following areas based on the foundations laid in our previous project. First, to build a merit to evaluate the unpredictability and to evaluate the proposed architecture using a merit of the unpredictability. Precisely, to measure the safe period for reseeding. Second, design a phase locked loop (PLL) embedded with a analog to digital mechanism to generate true random seed for digital chaos circuit. Third, integrate the chaotic seed to MRG in order to enhance the unpredictability rate and to guarantee good statistic characteristics.

Project ID:PB10308-3451
External Project ID:MOST103-2221-E182-072