I can show that a process $X(t)$ is Wide Sense stationary (WSS) by showing that $E[X(t)]$ is constant and that its autocorrelation function is in function of $\tau=t_1-t_2$, that is, $R_X(t+\tau,t)=R_X(\tau)$.

My question is:

If $E[X(t)]$ is constant and $R_X(t+\tau,t)=0$ can I say the process is WSS? What does $R_X(t+\tau,t)=0$ mean? Can I say $R_X(t+\tau,t)=0=R_X(\tau)$ and therefore is WSS?

  • $\begingroup$ by the way, almost certain that you also need the restriction that either a) the ACF is bounded or b) the distribution of any $X(t)$ has a bounded variance. That's mathematical nitpicking, though. (Imagine a process that is uncorrelated, zero-mean, cauchy-distributed: it's not WSS.) $\endgroup$ – Marcus Müller Jun 17 '18 at 15:38

If E[X(t)] is constant and RX(t+τ,t)=0 can I say the process is WSS?

Can I say RX(t+τ,t)=0=RX(τ) and therefore is WSS?

Two times the same question. It fulfills the definition (as you noticed yourself), so why even ask? yes.

What does RX(t+τ,t)=0 mean?

It means the process is uncorrelated ($R_X(\tau \ne 0)=0$), but also that its variance is 0 ($R_X(0) = \sigma_X^2 =0$). There's only one process that fulfills that:

$X(t) = 0$

not a really exciting process, is it?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.