Efficient and cryptographically secure pseudorandom number generators based on chains of hybrid cellular automata maps

In this paper we consider several solutions for designing cryptographically good pseudo random number generators (PRNG) with low complexity implementations. Our solutions are based on hybrid cellular automata (HCA) maps. The first solution (ensuring maximal throughput) is based on creating chains of HCA maps, i.e. the nonlinear map is changed dynamically being controlled by another HCA map within a chain. The second solution is simpler but somehow reduces the throughput and it is based on the down-sampling of a single HCA output with a factor d. It is shown that both solutions pass all statistical tests of NIST.