A nonlinear PRNG using digitized logistic map with self-reseeding method

This paper proposes a self-reseeding pseudo RNG (PRNG) based on digitized logistic map to introduce small perturbations periodically so that the low-period can be removed and the randomness can be enhanced. The strategy and the optimization of selecting reseeding period with corresponding fixed pattern are demonstrated. The resultant PRNG generates random sequence fulfilling all the randomness requirements in proportion of NIST SP 800-22 test suite, and the average period can be more than 65 folds. Furthermore, the proposed PRNG is validated by a hardware implementation using TSMC 0.18 μm CMOS process with throughput faster than 250 Mbit/s.