Asymptotically optimal codebooks derived from generalised bent functions

02/15/2020 ∙ by Qiuyan Wang, et al. ∙ NetEase, Inc 0

Codebooks are required to have small inner-product correlation in many practical applications, such as direct spread code division multiple access communications, space-time codes and compressed sensing. In general, it is difficult to construct optimal codebooks. In this paper, two kinds of codebooks are presented and proved to optimally optimal with respect to the welch bound. Additionally, the constructed codebooks in this paper have new parameters.

READ FULL TEXT VIEW PDF
POST COMMENT

Comments

There are no comments yet.

Authors

page 1

page 2

page 3

page 4

This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.

1 Introduction

An codebook

is defined to be a set of unit-norm complex vectors

in over an alphabet A. Let

where denotes the Hermite transpose of vector . The maximum inner-product correction is a performance measure of a codebook in practical applications. In code division multiple access (CDMA) systems, one important problem is to minimize the codebook’s maximal cross-correlation amplitude .

For a given , it is usually desirable that is as large as possible and is as small as possible simultaneously. However, the parameters , and of a codebook have to satisfy the Welch bound [25]. That is, there is a tradeoff between the codeword length , the set size and . A codebook meeting the theoretical bound with equality is said to be optimal. Searching optimal codebooks has been an interesting research topic in recent years. Many classes of optimal codebooks has been constructed [26, 19, 1, 23, 3, 4, 5, 6, 7, 18].

It is worthwhile to point out that the constructed codebooks so far have restrictive parameters and . Hence, many researchers attempt to research asymptotically optimal codebooks, i.e., asymptotically meets the theoretical bound for sufficiently large . One important method to construct asymptotically optimal codebooks is from difference sets of finite abelian gruops which is developed by Ding and Feng [3, 4]. In [3, 4], several series of asymptotically codebooks were constructed by using almost difference sets. In [8, 27, 2], asymptotically optimal codebooks were presented by binary row selection sequences. Character sums over finite fields are considered to be useful tools for the design of asymptotically codebooks. Recently, in [9, 10], Heng et al. obtained two new constructions of infinitely many codebooks by Jacobi sums and their generalizations. In [15, 16], Luo and Cao defined a new character sum called hyper Eisenstein sum and presented two constructions of infinitely many new codebooks achieving the Wech bound. In [24], Tian presented two constructions of codebooks with additive characters over finite fields.

Bent functions are a class of Boolean functions and have important applications in cryptography, code theory and sequences for communications. In cryptography, bent vectorial functions can be used as substitution boxes in block ciphers (ensuring confusion, as explained by Shannon [20]. In code theory, they are useful for constructing error correcting codes (Kerdock codes) [17]. In sequences, they permit to construct sequences with low correlation [11].

The objective of this paper is to present two constructions of codebooks using generalised bent functions over a ring of integers modulo . The presented two kinds of codebooks have properties: (1) they are asymptotically optimal with respect to the Welch bound; (2) the parameters of these codebooks are new and flexible. As a comparison with the known ones, our codebooks are listed in Table 1.

References Parameters Constraints
[8]

is an odd prime

[28] is an odd prime power
[13] is a power of a prime
[13] is a prime power
[27] is an odd prime
[29]
is a prime power
[9]
is a power of a prime,
[9]
,
is a prime power and
[14]
,
, ,
is a prime power
[14]
, ,
is a prime power
[14]
, ,
is a prime power
[24] is a prime power
[24] is a prime power
Theorem 3.1
is an integer and
is the smallest
prime factor of
Theorem 4.1
is an integer and
is the smallest
prime factor of
Table 1: The parameters of codebooks asymptotically meeting the Welch bound

This paper is organized as follows. In Section 2, we briefly recall some definitions and notations which will be needed in our discussion. In Section 3 and Section 4, we present our constructions of codebooks and prove they are asymptotically optimal with respect to the Welch bound. In Section 5, concluding remarks of this paper is given.

2 Preliminaries

In this section, we present some notations and preliminaries which are needed for the proof of the main results. Firstly, a useful lemma is given in the following.

Lemma 2.1 (Linear Congruence Theorem, [22]).

Let , , and be integers with , and let .

  1. If , then the congruence has no solutions.

  2. If , then the congruence has exactly incongruent solutions.

The following is the well-known Welch bound on , and of a codebook .

Lemma 2.2.

[25] For any codebook with ,

Moreover, the equality holds if and only if for all pairs of with , it holds that

Next, we introduce the definition of generalised bent functions. Let be a positive integer and the ring of integers modulo . Assume that is a primitive -th root of unity. Denote by the -dimensional vector space over . A function mapping from to is termed a generalised Boolean function on variables. For a generalised Boolean function , if the complex Fourier coefficients

preserve unit magnitude for any , then is a generalised bent function.

Bent functions are a hot research topic due to their wide applications in cryptography, information theory and coding theory. Kumar et al. [12] introduced the definition of generalised bent functions from to and gave a class of generalised bent functions in the following lemma.

Lemma 2.3 ([12]).

Assume that is a positive integer. Let be an arbitrary permutation and an arbitrary function on . Then the function

is generalised bent, where .

For more details on bent functions and generalised bent functions, we refer readers to [21]. Inspired by the generalised bent functions given in Lemma 2.3, we propose two constructions of codebooks in the following two sections.

3 The first construction of asymptotically optimal codebooks

In this section, we present a construction of codebooks and show that the maximum inner-product correction of these codebooks asymptotically achieves the Welch bound. Before proposing our construction, we need to do some preparations.

Suppose that is an integer and is the smallest prime factor of . Denote the standard basis of the -dimensional Hilbert space by which is formed by vectors of length as follow:

In the following theorem, we give a new construction of infinite many codebooks and evaluate their maximum inner-product correction.

Theorem 3.1.

Let symbols be the same as above. For any , , define a unit-norm complex vector of length by

where and are permutations on . Let

(1)

Then the set is a codebook with .

Proof.

According to the definition of codebooks, we deduce that is consisted of codewords with length . In other words, is a codebook with parameters . Now we turn to the computation of . Let , be two distinct codewords. We distinguish among the following three cases to calculate the correlation of and .

(1) If , , it is easy to verify that .

(2) If and , it is obvious that .

(3) If , , write and , where . Then we deduce that

If , we obtain

where the second identity follows from the fact that .

If , then . By Lemma 2.1, the congruence has only one integer solution in . In this case, we have

where the last equality is derived from the fact that for any . Therefore, we obtain .

The analysis above shows that . This completes the proof of this theorem. ∎

The next theorem deals with the asymptotical optimality of the codebooks defined in Theorem 3.1.

Theorem 3.2.

Let symbols be the same as before. Then the maximum inner-product correction of the codebook defined in (1) asymptotically meets the Welch bound.

Proof.

From Theorem 3.1, we know is a codebook. Obviously, Note that the corresponding Welch bound of is

Then we obtain

Observe that

which implies that the codebook asymptotically meets the Welch bound. ∎

In Table 2, we list some examples of codebooks generated by Theorem 3.1. The numerical results indicate that the codebooks defined in Theorem 3.1 asymptotically achieve the Welch bound as increases, as predicted in Theorem 3.2.

5 1225 0.02857 0.02608 0.91293
13 48841 0.96362
17 243049 0.97183
31 3575881 0.98425
43 9320809 0.98857
61 1229410598 19829209 0.99190
73 3123614666 42211009 0.99322
83 4583692596 54567796 0.99403
97 11774065058 120143521 0.99488
Table 2: The parameters of the codebook in Theorem 3.1

4 The second construction of asymptotically optimal codebooks

In this section, we propose a construction of codebooks by slightly modifying the construction of codebooks in Section 3. In addition, we show that these codebooks are asymptotically optimal with respect to the Welch bound.

Let be an integer and the smallest prime factor of . Assume that

is a set consisted of all rows of the identity matrix

. That is to say, is the standard basis of the Hilbert space with dimension . Let and be permutations on . For any , we define a set by

(2)

where

Let

(3)

Then is a codebook with parameters and the the maximum inner-product correction of can be obtained in the following theorem.

Theorem 4.1.

Assume that is an integer and is the smallest prime factor of . Then defined by (3) is a codebook with .

Proof.

Let , be two distinct codewords in . Now we calculate the correlation of and by distinguishing among the following cases.

(1) If , , it is obvious that .

(2) If and , we have .

(3) If , , we assume that and , where . Then we have

If and , ‘we obtain

If and , it follows from the fact that that

If , then . According to Lemma 2.1, the congruence has only one integer solution in .

When , we can deduce that

When , we have

Hence, we get that

Summarizing the conclusions in the three cases above, we obtain

This completes the proof. ∎

Theorem 4.2.

Let symbols be the same as before. If , then the codebook is asymptotically optimal with respect to the Welch bound.

Proof.

For and , by Lemma 2.2 we have

Hence,

Consequently,

This completes the proof of this theorem. ∎

Table 3 presents some parameters of codebooks derived from Theorem 4.1. From Table 3, we can see is very close to as increases. This means that the codebooks defined in Theorem 4.1 are asymptotically optimal with respect to the Welch bound for large , as predicted in Theorem 4.2.

7 5852 0.013072 0.01224 0.93618
19 190532 0.97474
29 1150256 0.98321
41 4719756 0.98803
59 12949202 0.99163
67 1538770575 22624292 0.99262
79 3439533363 42987692 0.99373
89 6707573377 74520056 0.99443
101 12362193253 121187072 0.99509
Table 3: The parameters of the codebook in Theorem 4.1

5 Concluding remarks

Employing generalised bent functions, we presented two classes of codebooks and proved that these constructed codebooks nearly meet the Welch bound. As a comparison, the parameters of some known classes of asymptotically optimal codebooks with respect to the Welch bound and the new ones are listed in Table 1. Obviously, the parameters of our classes of codebooks have not been covered in the literature.

References

  • [1] J. Conway, R. Harding, and N. Sloane, “Packing lines, planes, etc.: Packings in Grassmannian spaces,” Exp. Math., vol. 5, no. 2, pp. 139-159, 1996.
  • [2] X. Cao, W. Chou and X. Zhang, “More constructions of near optimal codebooks associated with binary sequences,” Adv. Math. Commun., vol. 11, no. 1, pp. 187-202, 2017.
  • [3] C. Ding, “Complex codebooks from combinatorial designs”, IEEE Trans. Inf. Theory, vol. 52, no. 9, pp. 4229-4235, 2006.
  • [4] C. Ding and T. Feng, “A generic construction of complex codebooks meeting the Welch bound,” IEEE Trans. Inf. Theory, vol. 53, no. 11, pp. 4245-4250, 2007.
  • [5] M. Fickus and D. Mixon, “Tables of the existence of equiangular tight frames,” arXiv:1504.00253v2, 2016.
  • [6] M. Fickus, D. Mixon and J. Jasper, “Equiangular tight frames from hyperovals,” IEEE Trans. Inf. Theory, vol. 62, no. 9, pp. 5225–5236, 2016.
  • [7] M. Fickus, J. Jasper, D. Mixon and J. Peterson, “Tremain equiangular tight frames,” arXiv:1602.03490v1, 2016.
  • [8] S. Hong, H. Park, T. Helleseth, and Y. Kim, “Near optimal partial Hadamard codebook construction using binary sequences obtained from quadratic residue mapping,” IEEE Trans. Inf. Theory, vol. 60, no. 6, pp. 3698-3705, 2014.
  • [9] Z. Heng, C. Ding and Q. Yue, “New constructions of asymptotically optimal codebooks with multiplicative characters,” IEEE Trans. Inf. Theory, vol. 63, no. 10, pp. 6179-6187, 2017.
  • [10] Z. Heng, “Nearly optimal codebooks based on generalized Jacobi sums,” Dis. Appl. Math., vol. 25, no. 11, pp. 227-240, 2018.
  • [11] M. G. Karpovsky, K. J. Kulikowski, Z. Wang,On-line self error detection with equal protection against all errors, Int. J. High. Reliab. Electron. Syst. Des., 2008.
  • [12] P. Kumar, R. Scholtz, L. Welch, “Generalized bent functions and their properties,” J. Combin. Theory Ser. A, vol. 40, no .1, pp. 90–107, 1985.
  • [13] C. Li, Q. Yue, Y. Huang, “Two families of nearly optimal codebooks,” Des. Codes Cryptogr., vol. 75, no. 1, pp. 43-57, 2015.
  • [14] G. Luo and X. Cao, “Two constructions of asymptotically optimal codebooks via the hyper Eisenstein sum,” IEEE Trans. Inf. Theory, vol. 63, no. 10, pp. 6179-6187, Oct. 2017.
  • [15] G. Luo and X. Cao, “New constructions of codebooks asymptotically achieving the Welch bound,” in Proc. IEEE Int. Symp. Inf. Theory, Vail, CO, USA, June 2018, pp. 2346-2349.
  • [16] G. Luo and X. Cao, “Two constructions of asymptotically optimal codebooks,” Crypt. Commun., vol. 11, pp. 825-838, 2019.
  • [17] F. J. MacWilliams, N. J. Sloane, The theory of error-correcting codes. North Holland, Amsterdam, 1977.
  • [18] F. Rahimi, “Covering graphs and equiangular tight frames,” Ph.D. Thesis, University of Waterloo, Ontario, 2016 (available at http://hdl.handle.net/10012/10793).
  • [19] D. Sarwate, “Meeting the Welch bound with equality,” Sequences and their Applications, Springer London, pp. 79-102, 1999.
  • [20] C. E. Shannon, “A mathematical theory of communication,” Bell Syst. Tech. J. 27, 379–423, 1948.
  • [21] S. Mesnager. Bent Functions: Fundamentals and Results. Springer, Switzerland, 2016.
  • [22] J. H. Silverman, A friendly introduction to number theory, Upper Saddle River, NJ: Pearson Prentice Hall, 2006.
  • [23] T. Strohmer and R. Heath, “Grassmannian frames with applications to coding and communication,” Appl. Comput. Harmon. Anal., vol. 14, no. 3, pp. 257-275, 2003.
  • [24] L. Tian, Y, Li, T. Liu and C. Xu, “Constructions of Codebooks Asymptotically Achieving the Welch Bound with Additive Characters,” IEEE Signal Processing Letters, vol. 26, no. 4, pp. 622-626, 2019.
  • [25] L. Welch, “Lower bounds on the maximum cross correlation of signals,” IEEE Trans. Inf. Theory, vol. 20, no. 3, pp. 397-399, 1974.
  • [26] P. Xia, S. Zhou, and G. Giannakis, “Achieving the Welch bound with difference sets,” IEEE Trans. Inf. Theory, vol. 51, no. 5, pp. 1900-1907, 2005.
  • [27] N. Yu, “A construction of codebooks associated with binary sequences,” IEEE Trans. Inf. Theory, vol. 58, no. 8, pp. 5522-5533, 2012.
  • [28] A. Zhang and K. Feng, “Two classes of codebooks nearly meeting the Welch bound,” IEEE Trans. Inf. Theory, vol. 58, no. 4, pp. 2507-2511, 2012.
  • [29] Z. Zhou, X. Tang, “New nearly optimal codebooks from relative difference sets,” Adv. Math. Commun., vol. 5, no. 3, pp. 521-527, 2011.