Документ взят из кэша поисковой машины. Адрес оригинального документа : http://chaos.phys.msu.ru/loskutov/PDF/TechPhLt10_04_engl.pdf
Дата изменения: Thu Feb 5 12:00:56 2009
Дата индексирования: Mon Oct 1 20:05:05 2012
Кодировка:
Technical Physics Letters, Vol. 30, No. 10, 2004, pp. 843­845. Translated from Pis'ma v Zhurnal Tekhnicheskooe Fiziki, Vol. 30, No. 20, 2004, pp. 1­7. Original Russian Text Copyright © 2004 by Loskutov, Rybalko, Churaev.

Data Encoding by Stabilization of Dynamical System Cycles
A. Yu. Loskutov*, S. D. Rybalko, and A. A. Churaev
Moscow State University, Moscow, 119899 Russia Bauman State Technical University, Moscow, 107005 Russia * e-mail: Loskutov@moldyn.phys.msu.ru
Received April 1, 2004

Abstract--A new method is proposed for masking transferred data with the aid of chaotic maps. The cryptographic stability is analyzed by the method of total probing. A correlation analysis of the obtained codes is performed and predictability of the code sequence is evaluated. A network application is developed, which allows legal users exchange messages protected by the proposed method. © 2004 MAIK "Nauka / Interperiodica".

In the present-day stage of development of the communication technologies, the problem of protecting information is among most important. In this Letter, we propose an original method of data encoding, which makes use of the possibility of stabilizing the cycles of chaotic maps. This possibility is based on the wellknown fact of the theory of dynamical systems [1­3] (see also [4, 5] and references therein): there exist periodic perturbations of the chaotic dynamical systems belonging to rather general types, which lead to stabilization of the cycle with a given period. Although data encoding by means of chaotic systems is now very popular (see, e.g., [6­11] and references therein), the proposed method is advantageous in allowing a network application to be developed for exchanging messages without preliminary synchronization of the transmitter and receiver (which is usually necessary in other approaches). Moreover, programs to be developed in the nearest future will allow sound messages to be encoded as well. In order to explain the proposed principle of data encoding, we will first describe the main theoretical result concerning the stabilization of cycles. Consider a map of some region M and Rj into itself: Ta : x f ( x, a ) , (1)

^ a = (a1, ..., a), ai aj, 1 i, j , i j, a1, ..., a A}, AR


Let us introduce a submanifold Ac A corresponding to only the chaotic behavior of map (1). In some papers (see, e.g., [2, 12­12]), it was proved that, for j = ^ 1 and j = 2, there exist perturbations a = (a1, a2, ..., a) ^ such that, for a Ac (or g(a) Ac), a perturbed map will be regular with a stable cycle of period t = n. Moreover, the following exact result is valid for onedimensional maps (j = 1) [5]. Let a map Ta : x f (x, a), x M, a A to obey the conditions: (i) there exists a submanifold M such that, for any x1, x2 , there can be found a* A for which f (x1, a*) = x2 and (ii) there exists a critical point xc such that f (x, a)/ x x = xc Dxf (xc, a) = 0 for any a A. Then, for any x2, x3, ..., x , there can be found x1 and a1, a2, ..., a such that the cycle (x1, x2, ..., ^ x) will be a stable cycle of perturbed map Ta for a = (a1, ..., a). For data encoding, it is necessary to develop a method for evaluating the permissible noise level (see [15]). This can be readily done as follows [5]. Let ^ the perturbed map Ta for a = (a1, a2, ..., a) to have a stable cycle of period , p = (x1, x2, ..., x). Then, provided that a i a = 1/ A a LS
­1 x

where a is a parameters from the manifold of possible values A R, x = {x1, ..., xj}, and f = {f1, ..., fj}. Let us introduce the concept of parametric perturbation. The most natural way of doing this is to define a map with respect to parameter a, which would determine its value at each moment of time, G : A A, a g(a). A perturbation will be called periodic with a period of , provided that the function g(a) is defined only in points a1, ..., a in the following manner: ai + 1 = g(ai), i = 1, ..., ­ 1; and a1 = g(a). In this case, the set of pertur-

1063-7850/04/3010-0843$26.00 © 2004 MAIK "Nauka / Interperiodica"


@

bations with period can be brought into correspon^ dence with manifold A = { a A A ... A :


i=1



i S x ,


844

Loskutov et al.

Table 1. The principle of encoding symbols and letter sequences for secure data transmission Y Symbol x 0, ..., x n 1, n 2, n 3 ASCII codes

Cycles of periods n i + 2 --
n 1 + 2,

..., x

( n 1 + 2 ) + n 2 + 2,

-- ..., x
( n1 + 2 + n2 + 2 ) + n3 + 2

Manifold of dynamical variables
n3 + 2

{ a 1, ..., a n1 + 2, { b 1, ..., b n2 + 2 }, { c 1, ..., c Manifold of parameters

}

a 1, ..., c n3 + 2 Sequence of numbers

(where i = 1, 2, ..., ; Sa = Вmax|Daf (x, a)|; L =
x, a

Вmax | D x f (x, a)|; and Sx = max |Dxf (x, a)|, this map also В
2

has a stable cycle, p' = (x1 + x1, x2 + x2, ..., x + x) ^ of period for a' = (a1 + a1, a2 + a2, ..., a + a), where |xi| x = 1/L S
­1 x

x, a

x, a

.

In the first step of encoding, it is necessary to obtain the ASCII codes of all symbols involved in the text to be encoded. As is known, each symbol in the ASCII system corresponds to a unique triad of integers. For example, Latin letter "a" corresponds to the ASCII code 97 with the triad n1 = 0, n2 = 9, n3 = 7. Then, each member of a triad is interpreted as the period of a cycle inherent in a dynamical system. In order to avoid degenerate cycles (period 0) and stable cycles (period 1), we add 2 to each ni (i = 1, 2, 3). Now, using the chaotic properties of an applied map (or the random number generator), we create a sequence with a length equal to the sum of all ni (increased by 2) plus 1. The last element is used for beginning the count of cycle periods. The obtained sequence of random numbers is considered as the sequence of values of the dynamical variable x. For this sequence to bear information concerning the encoded symbols, we replace a part of elements by the values of critical points xc , that is, the points where f '(a, x) ) xc = 0. These points are separated by ni + 2 steps beginning with the first. Thus, the sequence consists of subsequences, the number of which is equal to the number of members in the sequence (ni + 2), that is, to the number of symbols in the coded text multiplied by three. The periods of cycle will be equal to ni + 2. Now let us calculate the values of the control param^ eter a = a1, ..., an , that is, determine the perturbation stabilizing the obtained sequence of cycles. This can be readily done by considering the inverse problem of determining the parameters from the form of the map.

^ For particular maps, perturbations a producing stabilization of the cycle of a given period form a certain region in the parametric space. This circumstance can be used for encoding repeated symbols by means of random selection of parameters from this region. The main steps of the data encoding protocol using the proposed method are presented in Table 1. The final sequence a1, ..., c n3 + 2 (representing parameters rather than the message) is sent to a transmitter, where all operations (with certain differences related to rounding) are performed in the reverse order (the method is symmetric). In order to justify the proposed method, iut is necessary to perform a statistical correlation analysis and evaluate the cryptographic stability [16]. The statistical analysis was performed using a sequence of 900 values of the control parameters, encoding a message consisting of 1000 Latin letter "o" symbols. The transmission of this symbol represents the most dangerous regime of operation of the proposed method, since the ASCII code of this symbol is 111 and the information about each "o" is contained in the three sequential cycles of period ni + 2 = 1 + 2 = 3, whose repetition is highly undesired. Satisfactory results obtained in this particular case will provide evidence of even greater reliability of the proposed method in the case of encoding other symbols. The statistical analysis gave the following results: correlation coefficient, r = 0.0077; regression equation, y = 4.9322905 + 0.00769911509x; the average value in the set, x = 4.96478169. Therefore, the proposed method of data encoding is highly reliable from the standpoint of correlation analysis and is capable of protecting data messages of considerable length. The main qualitative measures of cryptographic stability of an encrypting system are the laboriousness and reliability of the cryptographic analysis [17, 18]. We have evaluated the cryptographic stability of the proposed data encoding protocol by method of total probing, which consists in sequential random and equiprobable trial of N keys without repeats from manifold K.
Vol. 30 No. 10 2004

1

TECHNICAL PHYSICS LETTERS


DATA ENCODING BY STABILIZATION OF DYNAMICAL SYSTEM CYCLES Table 2. Results of evaluation of the laboriousness of decoding Byte/coefficient 1 2 3 4 5 K 227 251 275 299 2123 E,


845

t(E) 67 s 30 years 6 в 108 years 1016 years 1.5 в 1023 years

226 250 274 298 2122

Note: the left column indicates the number of bytes intended for encoding the control parameter; the right column shows the laboriousness of decoding converted into time assuming the computation speed to be equal to that of modern supercomputers.

Thus, the main advantages of the proposed method of data encoding are as follows: (i) the protocol can be implemented using a rather wide class of maps; (ii) the transmitted signal contains only information necessary for the further data processing, rather than the informative message as such; (iii) the dynamical system, which serves as the key for decoding, possesses chaotic properties; (iv) decoding process does not require preliminary synchronization of the transmitter and receiver; (v) the method is stable with respect to external noise; (vi) theoretically, the number of possible variants is infinite. As for practical realization, a network application has been developed, which allows legal users exchange messages protected by the proposed method. REFERENCES

The process of probing is terminated upon testing k keys, where k = j, 1 < j < N, j being the first key number for which the decoded text is considered substantially meaningful, or k = N if this event does not take place for j N. The decoded text is assessed for meaningfulness using the following hypotheses: H(0) for the open text and H(1) for a random text. In formulating a probabilistic model, the assessment procedure is characterized by the following errors: = P(H(1)/H(0)), the probability of rejecting a meaningful text, and = P(H(0)/H(1)), the probability of taking a meaningless text as meaningful. A model for calculation of the laboriousness of the cryptographic analysis can be formulated as E
,

1 ( k ) = --K


k=1

N

k(1 ­ )

k­1

N в ( N ­ k) + ----------- (k ­ 1) + (1 ­ ) + --- N (1 ­ ) K 1­ K ­ N + ------------- K

N­1


k=1

N

k(1 ­ )

k­1

N + N(1 ­ ) ,

where E, (k) is the mathematical expectation characterizing termination of the probing process after probing k keys and N is the number of probed keys. The results of calculations performed assuming errorless mechanism of taking decisions ( = 0, = 0) are summarized in Table 2. The reliability was evaluated using the relation P ( N , , ) = [ ( 1 ­ ) / K ]


t=1

N

(1 ­ )

t­1

.

Obviously, reliability of the method of total probing assuming errorless mechanism of taking decisions ( = 0, = 0) is P = 1. SPELL: 1. iut
TECHNICAL PHYSICS LETTERS Vol. 30 No. 10

1. V. V. Alekseev and A. Yu. Loskutov, Dokl. Akad. Nauk SSSR 293, 1346 (1987) [Sov. Phys. Dokl. 32, 270 (1987)]. 2. A. Yu. Loskutov and A. I. Shishmarev, Usp. Mat. Nauk 48, 169 (1993). 3. E. Ott, C. Grebogi, and J. A. Yorke, Phys. Rev. Lett. 64, 1196 (1990). 4. S. Boccaletti, C. Grebogi, Y.-C. Lai, et al., Phys. Rev. 329, 103 (2000). [Please, check this reference] 5. A. Loskutov, Mat. Model. 12, 314 (2001). 6. K. M. Cuomo and A. V. Oppenheim, Phys. Rev. Lett. 71, 65 (1993). 7. S. Hayes, C. Grebogi, E. Ott, and A. Mark, Phys. Rev. Lett. 73, 1781 (1994). 8. L. Kocarev and U. Parlitz, Phys. Rev. Lett. 74, 5028 (1995). 9. T. L. Carroll and L. M. Pecora, Chaos 9, 445 (1999). 10. M. P. Kennedy and G. KolumbАn, Signal Process. 80, 1307 (2000). 11. B. Fraser, P. Yu, and T. Lookman, Phys. Rev. E 66, 017 202 (2002). 12. A. Loskutov and A. I. Shishmarev, Chaos 4, 391 (1994). 13. A. N. Deryugin, A. Yu. Loskutov, and V. M. Tereshko, Teor. Mat. Fiz. 104, 507 (1995). 14. A. N. Deryugin, A. Loskutov, and V. M. Tereshko, Chaos, Solitons and Fractals 7 (10), 1 (1996). 15. M. Dolnik and E. M. Bollt, Chaos 8, 702 (1998). 16. Contemporary Cryptology: The Science of Information Integrity, Ed. by G. J. Simmons (IEEE Press, New York, 1992). 17. Theory and Practice of Providing the Information Security, Ed. by P. D. Zegzhda ("Yakhtsmen," Moscow, 1996). 18. Cryptography, Ed. by V. P. Sherstyuk and E. A. Primenko (SOLON-R, Moscow, 2002).

Translated by P. Pozdeev

2004