On a family of pseudorandom binary sequences (Q2567413)

The author analyses the following family of pseudorandom binary sequences: Let \(p\) be an odd prime, \(g\) a primitive root modulo \(p\), and denote by \(\text{ind}~n\) the discrete logarithm of \(n\) to the base \(g\), i.e., \(\text{ind}~n=j\) if \(n=g^j\) with \(1\leq j\leq p-1\). Let \(f(X)\) be a polynomial of degree \(k\) modulo \(p\). Then the finite sequence \(e_1,\ldots,e_{p-1}\) is defined by \[ e_n=\begin{cases} 1 & \text{if}\;1\leq \text{ind}~f(n)\leq (p-1)/2,\\ -1 & \text{if}\;(p+1)/2\leq \text{ind}~f(n)\leq p-1\;\text{or}\;p\mid f(n). \end{cases} \] In particular, he studies the well-distribution measure and the autocorrelation measure for a single sequence introduced by \textit{C. Mauduit} and \textit{A. Sárközy} [Acta Arith. 82, No. 4, 365--377 (1997; Zbl 0886.11048)], the symmetry measure introduced by the author [Ramanujan J. 8, No. 3, 289--302 (2004; Zbl 1062.11053)], and the complexity measure for the family of the above sequences introduced by \textit{R. Ahlswede, L. H. Khachatrian, C. Mauduit} and \textit{A. Sárközy} [Period. Math. Hungar. 46, 107--118 (2003; Zbl 1050.11069)]. The proofs are based on character sum techniques. Finally, some numerical calculations are given.