On the independence of Hartman sequences (Q1606064)

Given \(A\subseteq [0,1)^k\) and \(\vec{\alpha}=(\alpha_1,\ldots,\alpha_k) \in {\mathbb R}^k\), we define \(\chi_{A,\vec{\alpha}}:{\mathbb Z}\rightarrow \{0,1\}\) by setting \(\chi_{A,\vec{\alpha}}(n)= 1\) if and only if \((\{\alpha_1n\},\ldots,\{\alpha_kn\})\in A\), where \(\{\alpha\}\) denotes the fractional part of \(\alpha\), i.e. \(\alpha\) is considered as an element of the torus \(\mathbb R/\mathbb Z\). If the topological boundary of \(A\) has Haar measure 0, then \(\chi_{A,\vec{\alpha}}\) is called a Hartman sequence, which is a generalisation of Kronecker and Beatty sequences. In this article the author answers a question of Winkler by showing explicitly for which sets \(A\subseteq [0,1)^k\), \(B\subseteq [0,1)^l\) and vectors \(\vec{\alpha}\in {\mathbb R}^k\), \(\vec{\beta}\in {\mathbb R}^l\) we have \(\chi_{A,\vec{\alpha}}=\chi_{B,\vec{\beta}}\). The main tool of the proof is Weyl's theorem on uniform distribution.