I'm trying to solve the following problem:
Random process {$X_t,t\in T$} is weak stationary process and has a canonical representation $X_t=U_1\cos\lambda t+U_2\sin\lambda,\lambda\in\mathbb{R}$. Let $U_k,k=1,2$ have a normal distribution $N(0,\sigma^2)$. Determine the correlation function of a process.
I know that dispersion of $X_t$ and dispersion of $X_s$ is $\sigma^2$, but I don't know how to calculate $\operatorname{Cov}(X_s,X_t)$.
Can you help me?