Next Article in Journal
A Note on Double Conformable Laplace Transform Method and Singular One Dimensional Conformable Pseudohyperbolic Equations
Previous Article in Journal
Group Geometrical Axioms for Magic States of Quantum Computing
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Autocorrelation Values of Generalized Cyclotomic Sequences with Period pn+1

School of Mathematics, Northwest University, Xi’an 710127, China
*
Author to whom correspondence should be addressed.
Mathematics 2019, 7(10), 950; https://doi.org/10.3390/math7100950
Submission received: 17 September 2019 / Revised: 9 October 2019 / Accepted: 9 October 2019 / Published: 12 October 2019
(This article belongs to the Section Mathematics and Computer Science)

Abstract

:
Recently Edemskiy proposed a method for computing the linear complexity of generalized cyclotomic binary sequences of period p n + 1 , where p = d R + 1 is an odd prime, d , R are two non-negative integers, and n > 0 is a positive integer. In this paper we determine the exact values of autocorrelation of these sequences of period p n + 1 ( n 0 ) with special subsets. The method is based on certain identities involving character sums. Our results on the autocorrelation values include those of Legendre sequences, prime-square sequences, and prime cube sequences.

1. Introduction

Pseudorandom sequences with good randomness properties are widely applied in simulation, radar systems, spread-spectrum communication systems, ranging systems, software testing, global positioning systems, channel coding, code-division multiple-access (CDMA) systems, and stream ciphers [1,2,3,4,5]. If a sequence s = ( s 0 , s 1 , ) over a field F satisfies s n = s n + T , n 0 , then s is said to be T-periodic. For a binary sequence s over the binary field F 2 , if s n = 1 if and only if n D , then set D is called the characteristic set or support set of s . Two important tools for measuring the randomness properties of pseudorandom sequences are autocorrelation values and linear complexity. The periodic autocorrelation value C s ( τ ) of a T-periodic binary sequence s is defined by:
C s ( τ ) = n = 0 T 1 ( 1 ) s n + τ + s n ,
where 0 τ T 1 . The periodic autocorrelation value reflects global randomness. The linear complexity L ( s ) of the sequence s is the length of the shortest linear feedback shift register which can generate s . It is defined to be the least positive integer L satisfying:
s t = c 1 s t 1 + c 2 s t 2 + + c L s t L for all t L ,
where c 1 , , c L F .
The construction and randomness properties analysis of pseudorandom sequences are the core problem for pseudorandom sequences theory. Ding and Helleseth [6] introduced a generalized cyclotomy with respect to p 1 e 1 p 2 e 2 p t e t , and introduced a class of new binary sequences whose characteristic set is selected as { 0 } [ d n , d > 1 n d D 1 ( d ) ] . Several generalized cyclotomic sequences were constructed based on this generalized cyclotomy. Ding [7] obtained lower bounds on the linear complexity of generalized cyclotomic sequences with period p 2 . He also determined the exact values of autocorrelation of these sequences by using certain formulas for generalized cyclotomic numbers. Later Ding [8] calculated the linear complexity of these sequences. In contrast to [7], this time he obtained the exact linear complexity, and the results did not require any special requirement for the prime. Park, Hong, and Eun [9] found technical errors in [8], so they corrected the errors and re-established the main results on the linear complexity. Yan, Sun, and Xiao [10] studied new generalized cyclotomic binary sequences with respect to p 2 , which are a special case of those whose characteristic set are { 0 } [ d n , d > 1 n d D 1 ( d ) ] in [6]. Results indicate that these sequences possess high linear complexity. The exact values of autocorrelation are five-valued for p 1 ( mod 4 ) , and three-valued for p 3 ( mod 4 ) . The exact same results on the linear complexity and the exact values of autocorrelation of these new binary sequences were presented in [11,12], respectively. Kim, Jin, and Song [13] calculated the exact values of autocorrelation and linear complexity of prime cube sequences with period p 3 . Results show that the autocorrelation values of these prime cube sequences are seven-valued for p 1 ( mod 4 ) , and four-valued for p 3 ( mod 4 ) . The linear complexity of generalized cyclotomic sequences of period p n for any n > 0 is calculated in [14,15,16,17]. The autocorrelation values of those sequences of order 2 and period p n are calculated in [18]. In this paper the described results on autocorrelation values are a generalization of the known ones from [10,11,12,13,19]. In contrast to [11,12,13], we present a simpler proof by using certain identities involving character sums. The proof of the results on autocorrelation values in [11,12,13] are all based on generalized cyclotomic numbers. The method for computing the autocorrelation values of the binary sequences in [10] is based on their characteristic polynomials. The autocorrelation values of our theorem are entirely consistent with those in [18], but the described results in this paper do not require the restriction d = 2 . In addition, the parameters of our sequences are more complicated than those in [18], but the proof of our results is shorter and simpler.
In 2011 Edemskiy [20] proposed a method for computing the linear complexity of p n + 1 -periodic generalized cyclotomic binary sequences. For details, suppose g is a primitive root of p n + 1 , where p = d R + 1 is an odd prime, d , R are two non-negative integers, and n > 0 is a positive integer. Let D 0 ( p n + 1 ) = ( g d ) be a cyclic subgroup of the multiplicative group Z p n + 1 * . Define D i ( p n + 1 ) = g i D 0 ( p n + 1 ) , i = 0 , 1 , , d 1 , p m D i ( p n + 1 ) = p m a : a D i ( p n + 1 ) , and m = 0 , 1 , , n . Then:
Z p n + 1 * = i = 0 d 1 D i ( p n + 1 ) and Z p n + 1 = m = 0 n i = 0 d 1 p m D i ( p n + 1 ) { 0 } .
Edemskiy defined the binary sequence s with period p n + 1 as follows:
s i = 1 , if ( i mod p n + 1 ) C , 0 , if ( i mod p n + 1 ) C ,
where the characteristic set of s is selected as:
C = m = 0 n i I m p m D i ( p n + 1 ) { 0 } , I m 0 , 1 , , d 1 .
He proposed a method for computing the linear complexity of s and considered some special given I m . In this paper for even d we shall choose special subsets:
I 0 = I 1 = = I n = 1 , 3 , 5 , , d 1 ,
and compute the exact values of autocorrelation of this special generalized cyclotomic sequence with period p n + 1 ( n 0 ) .

2. Sums Involving Legendre Symbol

We need the following lemma from [21].
Lemma 1.
Let p > 2 be a prime, l Z , and · p denote the Legendre symbol modulo p. Then we have:
t = 0 p 1 t ( t + l ) p = p 1 , p l , 1 , p l ,
where pl indicates that p is a divisor of l.
Lemma 2.
Let · p denote the Legendre symbol modulo p. For 1 τ p n + 1 1 with gcd ( τ , p n + 1 ) = p m 0 , we have:
m 1 = 0 n m 2 = 0 n t = 0 p n + 1 1 gcd ( t , p n + 1 ) = p m 1 gcd ( t + τ , p n + 1 ) = p m 2 t p m 1 p t + τ p m 2 p = p n + 1 p n m 0 ( p + 1 ) .
Proof. 
It is convenient to divide the relations between m 1 and m 2 into three cases according as m 1 > m 2 or m 1 < m 2 or m 1 = m 2 . From properties of the Legendre symbols modulo p and complete residue systems, we deduce:
m 1 = 0 n m 2 = 0 n m 1 > m 2 t = 0 p n + 1 1 gcd ( t , p n + 1 ) = p m 1 gcd ( t + τ , p n + 1 ) = p m 2 t p m 1 p t + τ p m 2 p = m 1 = 0 n m 2 = 0 n m 1 > m 2 t = 0 p n + 1 m 1 1 gcd ( p m 1 t + τ , p n + 1 ) = p m 2 t p p m 1 t + τ p m 2 p = m 1 = 0 n m 2 = 0 n m 1 > m 2 t = 0 p n + 1 m 1 1 gcd ( p m 1 t + τ , p n + 1 ) = p m 2 t p τ p m 2 p = m 1 = 0 n m 2 = 0 n m 1 > m 2 p m 2 τ t = 0 p n + 1 m 1 1 t p τ p m 2 p = 0 ,
and,
m 1 = 0 n m 2 = 0 n m 1 < m 2 t = 0 p n + 1 1 gcd ( t , p n + 1 ) = p m 1 gcd ( t + τ , p n + 1 ) = p m 2 t p m 1 p t + τ p m 2 p = m 1 = 0 n m 2 = 0 n m 1 < m 2 t = 0 p n + 1 1 gcd ( t τ , p n + 1 ) = p m 1 gcd ( t , p n + 1 ) = p m 2 t τ p m 1 p t p m 2 p = m 1 = 0 n m 2 = 0 n m 1 < m 2 t = 0 p n + 1 m 2 1 gcd ( p m 2 t + τ , p n + 1 ) = p m 1 t p p m 2 t τ p m 1 p = m 1 = 0 n m 2 = 0 n m 1 < m 2 t = 0 p n + 1 m 2 1 gcd ( p m 2 t + τ , p n + 1 ) = p m 1 t p τ p m 1 p = m 1 = 0 n m 2 = 0 n m 1 < m 2 p m 1 τ t = 0 p n + 1 m 2 1 t p τ p m 1 p = 0 ,
and,
m 1 = 0 n m 2 = 0 n m 1 = m 2 t = 0 p n + 1 1 gcd ( t , p n + 1 ) = p m 1 gcd ( t + τ , p n + 1 ) = p m 2 t p m 1 p t + τ p m 2 p = m = 0 n t = 0 p n + 1 m 1 gcd ( p m t + τ , p n + 1 ) = p m t p p m t + τ p m p = m = 0 n p m τ t = 0 p n + 1 m 1 gcd ( t + τ p m , p ) = 1 t p t + τ p m p + m = 0 n p m + 1 τ t = 0 p n + 1 m 1 gcd ( t , p ) = 1 t p t + τ p m p = m = 0 n p m τ t = 0 p n + 1 m 1 t p t + τ p m p + m = 0 n p m + 1 τ p n m ( p 1 ) = m = 0 n p m τ ( p n m ) + m = 0 n p m + 1 τ p n m ( p 1 ) = p n , if m 0 = 0 , p n m 0 + p n m 0 + 1 ( p m 0 1 ) , if 1 m 0 n , = p n + 1 p n m 0 ( p + 1 ) ,
by Lemma 1. The combined results in these three cases yield:
m 1 = 0 n m 2 = 0 n t = 0 p n + 1 1 gcd ( t , p n + 1 ) = p m 1 gcd ( t + τ , p n + 1 ) = p m 2 t p m 1 p t + τ p m 2 p = m 1 = 0 n m 2 = 0 n m 1 > m 2 t = 0 p n + 1 1 gcd ( t , p n + 1 ) = p m 1 gcd ( t + τ , p n + 1 ) = p m 2 t p m 1 p t + τ p m 2 p + m 1 = 0 n m 2 = 0 n m 1 < m 2 t = 0 p n + 1 1 gcd ( t , p n + 1 ) = p m 1 gcd ( t + τ , p n + 1 ) = p m 2 t p m 1 p t + τ p m 2 p + m 1 = 0 n m 2 = 0 n m 1 = m 2 t = 0 p n + 1 1 gcd ( t , p n + 1 ) = p m 1 gcd ( t + τ , p n + 1 ) = p m 2 t p m 1 p t + τ p m 2 p = p n + 1 p n m 0 ( p + 1 ) .

3. Autocorrelation Values

The object of this section is to compute the autocorrelation values of s and the main results are stated as follows.
Theorem 1.
Let s be defined as in Equation (1) with I m = 1 , 3 , 5 , , d 1 for m = 0 , 1 , , n . For 0 τ p n + 1 1 , the autocorrelation values of s are:
C s ( τ ) = p n + 1 , if τ = 0 , p n + 1 p n m 0 ( p + 1 ) τ p m 0 p τ p m 0 p , if gcd ( τ , p n + 1 ) = p m 0 ,
where · p denotes the Legendre symbol modulo p.
Proof. 
For t Z , let gcd ( t , p n + 1 ) = p m , 0 m n and write t = p m t , gcd ( t , p ) = 1 . Observe that by C = m = 0 n i I m p m D i ( p n + 1 ) { 0 } and make use of the orthogonality relation for characters modulo p n + 1 , we have:
t C there exists i I m satisfying t p m D i ( p n + 1 ) there exists i I m satisfying t D i ( p n + 1 ) there exist i I m , 0 s p n R 1 satisfying t g d s + i ( mod p n + 1 ) 1 ϕ ( p n + 1 ) i I m s = 0 p n R 1 χ mod p n + 1 χ ( t ) χ ¯ ( g d s + i ) = 1 1 d χ mod p n + 1 χ ( g d ) = 1 χ ( t ) i I m χ ¯ ( g i ) = 1 ,
where the sum χ mod p n + 1 is over all multiplicative characters modulo p n + 1 . With the aid of Equation (3) and the definition of I m we now get:
( 1 ) s t = 2 d χ mod p n + 1 χ ( g d ) = 1 χ χ 0 i I m χ ¯ ( g i ) χ ( t ) = 2 d χ mod p n + 1 χ ( g d ) = 1 χ χ 0 i = 0 d 2 1 χ ¯ ( g 2 i + 1 ) χ ( t ) = χ mod p n + 1 χ ( g d ) = 1 χ ( g 2 ) = 1 χ χ 0 χ ¯ ( g ) χ ( t ) = t p ,
where · p is the Legendre symbol modulo p. Then for 0 t p n + 1 1 , we have:
( 1 ) s t = t p , t = p m t , 0 m n , gcd ( t , p ) = 1 , 1 , p n + 1 t .
For 1 τ p n + 1 1 with gcd ( τ , p n + 1 ) = p m 0 , by means of Equation (4) we have:
C s ( τ ) = t = 0 p n + 1 1 ( 1 ) s t + τ + s t = m 1 = 0 n m 2 = 0 n t = 0 p n + 1 1 gcd ( t , p n + 1 ) = p m 1 gcd ( t + τ , p n + 1 ) = p m 2 t p m 1 p t + τ p m 2 p m = 0 n t = 0 p n + 1 1 gcd ( t , p n + 1 ) = p m p n + 1 t + τ t p m p m = 0 n t = 0 p n + 1 1 p n + 1 t gcd ( t + τ , p n + 1 ) = p m t + τ p m p + t = 0 p n + 1 1 p n + 1 t p n + 1 t + τ 1 = m 1 = 0 n m 2 = 0 n t = 0 p n + 1 1 gcd ( t , p n + 1 ) = p m 1 gcd ( t + τ , p n + 1 ) = p m 2 t p m 1 p t + τ p m 2 p τ p m 0 p τ p m 0 p .
It follows from Lemma 2 that:
C s ( τ ) = p n + 1 , if τ = 0 , p n + 1 p n m 0 ( p + 1 ) τ p m 0 p τ p m 0 p , if gcd ( τ , p n + 1 ) = p m 0 ,
which proves Theorem 1. □
In the case of d = 2 , that is, if s is defined by the quadratic residue classes, then the autocorrelation values are entirely consistent with those when n = 0 in [19], n = 1 in [10,11,12], and n = 2 in [13]. As a consequence, we get the following two corollaries for two special cases.
Corollary 1.
If d = 2 , n = 1 , I m = 1 , m = 0 , 1 , that is:
s i = 1 , if ( i mod p 2 ) D 1 ( p 2 ) p D 1 ( p 2 ) { 0 } , 0 , if ( i mod p 2 ) D 0 ( p 2 ) p D 0 ( p 2 ) .
Then for p 1 ( mod 4 ) we have:
C s ( τ ) = p 2 , if τ = 0 , p 2 , if τ D 0 ( p 2 ) , p + 2 , if τ D 1 ( p 2 ) , p 2 p 3 , if τ p D 0 ( p 2 ) , p 2 p + 1 , if τ p D 1 ( p 2 ) ,
and for p 3 ( mod 4 ) we have:
C s ( τ ) = p 2 , if τ = 0 , p , if τ Z p 2 * , p 2 p 1 , if τ p Z p * .
Corollary 2.
If d = 2 , n = 2 , I m = 1 , m = 0 , 1 , 2 , that is:
s i = 1 , if ( i mod p 3 ) D 1 ( p 3 ) p D 1 ( p 3 ) p 2 D 1 ( p 3 ) { 0 } , 0 , if ( i mod p 2 ) D 0 ( p 3 ) p D 0 ( p 3 ) p 2 D 0 ( p 3 ) .
Then for p 1 ( mod 4 ) we have:
C s ( τ ) = p 3 , if τ = 0 , p 2 2 , if τ D 0 ( p 3 ) , p 2 + 2 , if τ D 1 ( p 3 ) , p 3 p 2 p 2 , if τ p D 0 ( p 3 ) , p 3 p 2 p + 2 , if τ p D 1 ( p 3 ) , p 3 p 3 , if τ p 2 D 0 ( p 3 ) , p 3 p + 1 , if τ p 2 D 1 ( p 3 ) ,
and for p 3 ( mod 4 ) we have:
C s ( τ ) = p 3 , if τ = 0 , p 2 , if τ D 0 ( p 3 ) D 1 ( p 3 ) , p 3 p 2 p , if τ p D 0 ( p 3 ) p D 1 ( p 3 ) , p 3 p 1 , if τ p 2 D 0 ( p 3 ) p 2 D 1 ( p 3 ) .

4. Conclusions

In this paper we computed the exact values of autocorrelation of generalized cyclotomic binary sequences of any order d and period p n + 1 ( n 0 ) . Theorem 1 included the results of the autocorrelation values of Legendre sequences, prime-square sequences, and prime cube sequences from [10,11,12,13,19]. The autocorrelation values of our theorem were entirely consistent with those in [18]. In contrast to [18], our main results did not need the restriction d = 2 , and the proof of our theorem was based on certain identities involving character sums while the proof in [18] used the generalized cyclotomic numbers.

Author Contributions

Conceptualization, X.C. and H.L.; methodology, X.C. and H.L.; writing—original draft preparation, X.C. and H.L.; writing—review and editing, X.C. and H.L.

Funding

This research was funded by National Natural Science Foundation of China grant number 11571277, and the Science and Technology Program of Shaanxi Province of China grant number 2016GY-080 and 2016GY-077.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Cusick, T.W.; Ding, C.; Renvall, A. Stream Ciphers and Number Theory; North-Holland Mathematical Library, North-Holland Publishing Co.: Amsterdam, The Netherlands, 1998. [Google Scholar]
  2. Ding, C. Autocorrelation values of generalized cyclotomic sequences of order two. IEEE Trans. Inf. Theory 1998, 44, 1699–1702. [Google Scholar] [CrossRef]
  3. Gong, X.; Zhang, B.; Feng, D.; Yan, T. Autocorrelation values of new generalized cyclotomic sequences of order six over Z p q . In Information Security and Cryptology; Lecture Notes in Computer Science; Springer: Berlin, Germany, 2014; Volume 8567, pp. 86–98. [Google Scholar]
  4. Li, S.; Chen, Z.; Fu, X.; Xiao, G. Autocorrelation values of new generalized cyclotomic sequences of order two and length pq. J. Comput. Sci. Technol. 2007, 22, 830–834. [Google Scholar] [CrossRef]
  5. Wang, Q. The linear complexity of some binary sequences with three-level autocorrelation. IEEE Trans. Inf. Theory 2010, 56, 4046–4052. [Google Scholar] [CrossRef]
  6. Ding, C.; Helleseth, T. New generalized cyclotomy and its applications. Finite Fields Their Appl. 1998, 4, 140–166. [Google Scholar] [CrossRef]
  7. Ding, C. Binary cyclotomic generators. In Fast Software Encryption; Lecture Notes in Computer Science; Preneel, B., Ed.; Springer: Berlin, Germany, 1995; Volume 1008, pp. 29–60. [Google Scholar]
  8. Ding, C. Linear complexity of some generalized cyclotomic sequences. Int. J. Algebra Comput. 1998, 8, 431–442. [Google Scholar] [CrossRef]
  9. Park, Y.-H.; Hong, D.; Eun, H. On the linear complexity of some generalized cyclotomic sequences. Int. J. Algebra Comput. 2004, 14, 431–439. [Google Scholar] [CrossRef]
  10. Yan, T.; Sun, R.; Xiao, G. Autocorrelation and linear complexity of the new generalized cyclotomic sequences. IEICE Trans. Fundam. Electron. Commun. Comput. Sci. 2007, E90-A, 857–864. [Google Scholar] [CrossRef]
  11. Bai, E.; Liu, X. Some notes on prime-square sequences. J. Comput. Sci. Technol. 2007, 22, 481–486. [Google Scholar] [CrossRef]
  12. Yan, T.; Huang, B.; Xiao, G. Cryptographic properties of some binary generalized cyclotomic sequences with the length p2. Inf. Sci. 2008, 178, 1078–1086. [Google Scholar] [CrossRef]
  13. Kim, Y.-J.; Jin, S.-Y.; Song, H.-Y. Linear complexity and autocorrelation of prime cube sequences. In Applied Algebra, Algebraic Algorithms, and Error-Correcting Codes; Lecture Notes in Computer Science; Springer: Berlin, Germany, 2007; Volume 4851, pp. 188–197. [Google Scholar]
  14. Du, X.; Yan, T.; Shi, Y. Linear complexity of generalized cyclotomic sequences with period pm. J. Electron. Inf. Technol. 2010, 32, 821–824. (In Chinese) [Google Scholar] [CrossRef]
  15. Kim, Y.-J.; Song, H.-Y. Linear complexity of prime n-square sequences. In Proceedings of the IEEE International Symposium on Information Theory, Toronto, ON, Canada, 6–11 July 2008; pp. 2405–2408. [Google Scholar]
  16. Shi, Y.; Du, X.; Yan, T.; Li, X. Study of linear complexity of generalized cyclotomic sequences with period pm. Comput. Eng. 2013, 39, 189–193. (In Chinese) [Google Scholar]
  17. Yan, T.; Li, S.; Xiao, G. On the linear complexity of generalized cyclotomic sequences with the period pm. Appl. Math. Lett. 2008, 21, 187–193. [Google Scholar] [CrossRef]
  18. Jin, S.-Y.; Kim, Y.-J.; Song, H.-Y. Autocorrelation of new generalized cyclotomic sequences of period pn. IEICE Trans. Fundam. Electron. Commun. Comput. Sci. 2010, E93-A, 2345–2348. [Google Scholar] [CrossRef]
  19. Golomb, S.W.; Gong, G. Sequence Design for Good Correlation; Cambridge University Press: Cambridge, UK, 2005. [Google Scholar]
  20. Edemskiy, V. About computation of the linear complexity of generalized cyclotomic sequences with period pn+1. Des. Codes Cryptogr. 2011, 61, 251–260. [Google Scholar] [CrossRef]
  21. Mullen, G.L.; Panario, D. Handbook of Finite Fields; Chapman and Hall/CRC: New York, NY, USA, 2013. [Google Scholar] [CrossRef]

Share and Cite

MDPI and ACS Style

Chen, X.; Liu, H. Autocorrelation Values of Generalized Cyclotomic Sequences with Period pn+1. Mathematics 2019, 7, 950. https://doi.org/10.3390/math7100950

AMA Style

Chen X, Liu H. Autocorrelation Values of Generalized Cyclotomic Sequences with Period pn+1. Mathematics. 2019; 7(10):950. https://doi.org/10.3390/math7100950

Chicago/Turabian Style

Chen, Xiaolin, and Huaning Liu. 2019. "Autocorrelation Values of Generalized Cyclotomic Sequences with Period pn+1" Mathematics 7, no. 10: 950. https://doi.org/10.3390/math7100950

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop