English

Secrecy Gain of Formally Unimodular Lattices from Codes over the Integers Modulo 4

Information Theory 2024-02-14 v3 math.IT

Abstract

Recently, a design criterion depending on a lattice's volume and theta series, called the secrecy gain, was proposed to quantify the secrecy-goodness of the applied lattice code for the Gaussian wiretap channel. To address the secrecy gain of Construction A4\text{A}_4 lattices from formally self-dual Z4\mathbb{Z}_4-linear codes, i.e., codes for which the symmetrized weight enumerator (swe) coincides with the swe of its dual, we present new constructions of Z4\mathbb{Z}_4-linear codes which are formally self-dual with respect to the swe. For even lengths, formally self-dual Z4\mathbb{Z}_4-linear codes are constructed from nested binary codes and double circulant matrices. For odd lengths, a novel construction called odd extension from double circulant codes is proposed. Moreover, the concepts of Type I/II formally self-dual codes/unimodular lattices are introduced. Next, we derive the theta series of the formally unimodular lattices obtained by Construction A4\text{A}_4 from formally self-dual Z4\mathbb{Z}_4-linear codes and describe a universal approach to determine their secrecy gains. The secrecy gain of Construction A4\text{A}_4 formally unimodular lattices obtained from formally self-dual Z4\mathbb{Z}_4-linear codes is investigated, both for even and odd dimensions. Numerical evidence shows that for some parameters, Construction A4\text{A}_4 lattices can achieve a higher secrecy gain than the best-known formally unimodular lattices from the literature. Results concerning the flatness factor, another security criterion widely considered in the Gaussian wiretap channel, are also discussed.

Keywords

Cite

@article{arxiv.2303.08083,
  title  = {Secrecy Gain of Formally Unimodular Lattices from Codes over the Integers Modulo 4},
  author = {Maiara F. Bollauf and Hsuan-Yin Lin and Øyvind Ytrehus},
  journal= {arXiv preprint arXiv:2303.08083},
  year   = {2024}
}

Comments

arXiv admin note: text overlap with arXiv:2202.09236; Paper accepted for publication in IEEE Transactions on Information Theory