Physics Letters A 373 (2009) 653–660
Contents lists available at ScienceDirect
Physics Letters A
www.elsevier.com/locate/pla
Discrete asymptotic deterministic randomness for the generation
of pseudorandom bits
Kai Wang
∗
, Wenjiang Pei, Xubo Hou, Song Hong, Zhenya He
Department of Radio Engineering, Southeast University, Nanjing, China
article info abstract
Article history:
Received 13 August 2008
Received in revised form 28 August 2008
Accepted 8 December 2008
Available online 25 December 2008
Communicated by A.R. Bishop
PACS:
05.45.+b
Keywords:
Discrete asymptotic deterministic
randomness
Multi-value correspondence
PRBG
The deterministic randomness not only can become a dominant approach in exploring the relation-
ship between chaos and randomness, but also can be associated with some famous number theoretical
concepts and open problems in number theory. Compared with chaotic sequences, asymptotic determin-
istic randomness sequences have the characteristic of multi-value correspondence, which makes those
sequences unpredictable in short steps. In this Letter, we will propose the definition of the discrete
asymptotic deterministic randomness, and then analyze the dynamical characteristics such as maximum-
period and multi-value correspondence. Referring to the NIST80 0-22 statistical test suite, we will present
and discuss two examples of PRBGs based on DADR, from the point of view of FPGA design and random-
ness quality.
© 2008 Elsevier B.V. All rights reserved.
1. Introduction
Chaos is a random-like phenomenon generated by deterministic systems. Different from randomness sequences, the sequences gener-
ated by chaotic system are predictable in the short term, because the dynamical systems are determined. Today, chaos plays a key role
in the study of the relation between randomness and determinacy. Many works have focused on the construction of nonlinear dynamical
models which can generate unpredictable sequences in the short term [1–4].Refs.[5,6] are devoted to the analysis of the application of a
chaotic piecewise-linear map as Pseudorandom bit generators (PRBGs) and mathematically analyze the information generation process of
a class of piecewise linear 1D maps. Effective PRBGs are obtained by means of a chaotic system based on a pipeline analog-to-digital con-
verter [7]. The discretized chaos based PRBGs of low complexity are analyzed to evaluate its suitability for the integrated implementation
[8–10].
Furthermore, a remarkable work is proposed by J.A. Gonzalez, whose corresponding researches have discussed dynamics of the general
function: x
n
= sin
2
(πθ z
n
) when z is a relative prime fraction number. The sequence produced by this function is unpredictable in the short
term and has the characteristic of multi-value correspondence [11–14]. This phenomenon is named as “deterministic randomness (DR)”, in
order to differentiate it from chaos and randomness. Obviously, DR not only can become a dominant approach in exploring the relationship
between chaos and randomness, but also can be associated with some famous number theoretical concepts and open problems in number
theory [15]. Further studies show that we can construct generalized asymptotic deterministic randomness (ADR) systems by utilizing the
piecewise linear/nonlinear map and the noninvertible nonlinear transform [16,17], and in nature, this transformation is also similar to the
one-way function of cryptography [18].
Chaos in systems with discrete phase space can be calle d pseudo-chaos, quantum-chaos or, more generally, discrete chaos [19–24].
Refs. [19–21] propose a definition of the discrete Lyapunov exponent and discrete entropy. Furthermore, Refs. [22,23] utilize the discrete
Lyapunov exponent in chaotic cryptology and chaotic cryptanalysis. The maximum-period of digitized Renyi maps have been discussed in
Ref. [24]. The discrete asymptotic deterministic randomness (DADR) systems are constructed by utilizing the piecewise linear/nonlinear
map and the noninvertible nonlinear transform in discrete phase space. As a result, theoretical results concerning discrete chaos are
helpful for studying DADR. In this Letter, we will propose the definition of DADR, and then analyze the dynamical characteristics such
*
Corresponding author. Tel.: +86 025 83795996; fax: +86 025 83795996.
E-mail address: kaiwang@seu.edu.cn (K. Wang).
0375-9601/$ – see front matter
© 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.physleta.2008.12.037