Stationarity of independent sequences (Q1892108)

Let \(A(t) = \sum_{n \in Z} A_ n \mu_ n (t)\) be a real random function, such that \(\{A_ n\}\) are mutually independent, identically distributed random variables, \(\mu_ n(n) = 1\), \(\mu_ n(k) = 0\), \(k \neq n\), \(\sum_{n} \mu^ 2_ n(t) < \infty\), \(t \in \mathbb{R}\), \(E(A_ n) = 0\), \(E(A^ 2_ n) = 1\). Then a necessary and sufficient condition for \(A(t)\) to be stationary is that: \((\text{P}_ 1)\) \(\mu_ n(t) = \mu_ 0(t-n)\), \(\forall t \in \mathbb{R}\), \(n \in \mathbb{Z}\); \((\text{P}_ 2)\) \(\exists s(\omega)\), taking values 0 and 1 such that \(\mu_ 0(t) = {1\over 2\pi} \int^ \infty_{-\infty} e^{i \omega t} s(\omega) d \omega\), \(\sum_{k \in \mathbb{Z}} s(\omega + 2 k \pi) = 1\), \(\forall \omega \in \mathbb{R}\); \((\text{P}_ 3)\) \(A_ 0\) is normal, are satisfied.