A note on Barker sequences of even length

TL;DR

This paper investigates the properties of Barker sequences of even length, providing new symmetry insights and proving that no Barker sequence of even length greater than 4 exists under certain autocorrelation conditions.

Contribution

It introduces a weaker symmetry property for even-length Barker sequences and proves their nonexistence for sequences with specific autocorrelation patterns using elementary methods.

Findings

No Barker sequence of even length n>4 with equal autocorrelations at odd lags exists.

Established a weaker symmetry property for even-length Barker sequences.

Provided elementary proof techniques for nonexistence results.

Abstract

A Barker sequence is a binary sequence for which all nontrivial aperiodic autocorrelations are either 0, 1 or -1. The only known Barker sequences have length 2, 3, 4, 5, 7, 11 or 13. It is an old conjecture that no longer Barker sequences exist and in fact, there is an overwhelming evidence for this conjecture. For binary sequences of odd length, this conjecture is known to be true, whereas for even length it is still open, whether a Barker sequence of even length greater 4 exists. Similar to the well-known fact that a Barker sequence of odd length is necessarily skew-symmetric, we show that in the case of even length there is also a form of symmetry albeit weaker. In order to exploit this symmetry, we derive different formulas for the calculation of the aperiodic correlation. We prove by using only elementary methods that there is no Barker sequence of even length n>4 with $C_{1}=C_{3}=\cdots=C_{\frac{n}{2}-1}$, where $C_{k}$ denotes the $k$th aperiodic autocorrelation of the sequence.