On the FPGA implementation of random number generators from chaotic maps

Abstract

In this paper we describe the characteristics of four chaotic maps that are simulated by MATLAB to evaluate their entropy, bifurcation diagrams and Lyapunov exponent. From, the bifurcation diagrams, we choose values to search improved binary sequences using computer arithmetic of 32 bits in 2's complement and fixed point formats. In this manner, we choose the chaotic map with good entropy and positive Lyapunov exponent to implement a random number generator (RNG) into an FPGA. Finally, we show experimental results for the RNG that pass NIST 800-22 tests.