Nonlinear \(p\)-ary sequences (Q1403314)

The authors construct for every odd prime \(p\) and every integer \(m\geq 1\) a family \(F_{p,m}\) of pairwise cyclically non-equivalent and highly nonlinear \(p\)-ary sequences with the following parameters: period of sequences \(T= p(p^m- 1)\), cardinality of the family \(M= p^{m-1}\), maximum correlation \(\theta_{\max}\leq (p -1)p^{m/2+1}\). First, the authors construct the image of the smaller Kerdock code \(k_{p^2,m}\) over \(\mathbb{Z}_{p^2}\) (defined by its parity-check polynomial) by the Nechaev-Gray map \(\Phi\). Then in order to obtain sequences over \(\mathbb{Z}_p\) from those codes words, they use the standard additive character tool that maps an element of \(\mathbb{Z}_{p^2}\) into \(\mathbb{Z}_p\). The parameter \(T\) is obviously derived from the construction itself and the parameter \(M\) is calculated using the cyclicity of \(\Phi(k_{p^2,m})\). To prove the bound on the maximum correlation \(\theta_{\max}\) the authors need to establish a result on the trace representation (in terms of cyclotomic fields) of the Gray map image of an element of \(\mathbb{Z}^n_{p^2}\) which allows them to bound the modulus of each character sum (coming from the calculus of correlation). After parametrizing the \(p\)-adic expansion of the words of \(k_{p^2,m}\) by nonlinear functions, the authors eventually prove that \(\Phi(k_{p^2,m})\) is not \(\mathbb{F}_p\)-linear. The citations are correctly used and the results are well presented. The goal of this work appears very intuitive. From a multiple access communications systems point of view, this paper is a good contribution in the study of \(q\)-ary periodic sequences with particular correlation properties, and also from a cryptographic point of view as it allows highly nonlinear sequences to be constructed from cyclic codes.