Abstract:
In this paper the limit distribution for sums
$$
S_n=\sum_{k=1}^nf_{nk}(\xi_k^{(n)})
$$
is considered, where $f_{nk}(x)$ are measurable functions,
$$
\xi_k^{(n)}=\xi\biggl(\frac knT\biggr)-\xi\biggl(\frac{k-1}nT\biggr),\quad k=1,2,\dots,n,
$$
and $\xi(t)$, $t\in[0,T]$ is a stationary real-valued Gaussian process. Conditions are obtained under which the distributions of $S_k$ converge to a Gaussian and degenerate law.