Layered cellular automata for pseudorandom number generation

Syn Kiat Tan*, Sheng Uei Guan

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)

Abstract

The proposed Layered Cellular Automata (L-LCA), which comprises of a main CA with L additional layers of memory registers, has simple local interconnections and high operating speed. The time-varying L-LCA transformation at each clock can be reduced to a single transformation in the set {A f| = 1, 2,..., 2n -1} formed by the transformation matrix A of a maximum length Cellular Automata (CA), and the entire transformation sequence for a single period can be obtained. The analysis for the period characteristics of state sequences is simplified by analyzing representative transformation sequences determined by the phase difference between the initial states for each layer. The L-LCA model can be extended by adding more layers of memory or through the use of a larger main CA based on widely available maximum length CA. Several L-LCA (L = 1, 2, 3, 4) with 10- to 48-bit main CA are subjected to the DIEHARD test suite and better results are obtained over other CA designs reported in the literature. The experiments are repeated using the well-known nonlinear functions f30 and f45 in place of the linear function f204 used in the L-LCA. Linear complexity is significantly increased when f30 or f45 is used.

Original languageEnglish
Pages (from-to)217-234
Number of pages18
JournalInternational Journal of Modern Physics C
Volume18
Issue number2
DOIs
Publication statusPublished - Feb 2007
Externally publishedYes

Keywords

  • Cellular automata
  • Programmable cellular automata
  • Pseudorandom number generation

Fingerprint

Dive into the research topics of 'Layered cellular automata for pseudorandom number generation'. Together they form a unique fingerprint.

Cite this