Quaternary Golay sequence pairs. II: Odd length (Q2430711)

A quaternary Golay sequence pair of length \(s\) is a pair of sequences \(A, B\) of length \(s\) over \(\mathbb Z_4\) such that \(C_A(u) + C_B(u) = 0\) for all \(0<u<s\), where \(C_A(u)\) denotes the aperiodic autocorrelation \(C_A(u) := \sum_{j=0}^{s-1-u}i^{a[j]-a[j+u]}\) for an integer \(0\leq u\leq s-1\), when \(a[j]\) denotes the \(j\)-th term of the sequence \(A\). The authors show that one can obtain quaternary Golay sequence pairs of length \(s\) from a Barker sequence of length \(s\) when \(s\equiv 5 \bmod 8\). They further show that this explains the origin of all quaternary Golay sequence pairs of lengths \(5\) and \(13\). As is well known, there are only Barker sequences of odd length \(s\) for \(s=3,5,7,11,13\) (see [\textit{R. Turyn} and \textit{J. Storer}, Proc. Am. Math. Soc. 12, 394--399 (1961; Zbl 0116.01006)]) and it is even conjectured that there is no Barker sequence at all for \(s>13\). Therefore this result cannot serve to obtain quaternary Golay sequence pairs of odd length \(s>13\). As seen in Part I [the authors, Des. Codes Cryptography 59, No. 1--3, 131--146 (2011; Zbl 1233.94013)], there is in fact no quaternary Golay sequence pair of odd length \(s < 26\) and \(s \neq 3,5,11,13\). Accordingly, the authors give, among others, the interesting open questions: Is there a connection between quaternary Golay sequence pairs of lengths \(3\) and \(11\) and a Barker sequence of length \(3\) and \(11\)? Does there exist a quaternary Golay sequence pair of odd length greater than \(13\)?