Sang-Ho Shin and Kee-Young Yoo
Cellular automata, Pseudorandom number generator, Cryptography, Statistical randomness test
The linear neighborhood relationships in typical CA with 2-state and 3-neighborhood provide a high randomness quality. However, these relationships can cause the O problem which is the occurrence of similar events that are separated by different events. To reduce the O problem, this paper proposes a cellular automata evolution with 2-state and selective 4-neighborhood relationship. A selective 4-neighborhood relationship method is that the three CA cells at tth time step in four CA cell cases at (t-1)th time step in order to evolve the next state are chosen. To analyze the proposed scheme, the ENT and DIEHARD test suites are utilized in experiments. In experimental results, the randomness quality of the proposed PRNG was superior to the typical CA. Also, the occurrence of O problem was reduced in the proposed scheme.
Important Links:
Go Back
