1
$\begingroup$

Given

$$ Y(t) = A X(t) \cos(\omega t + \phi) $$

with $X(t)$ is zero-mean WSS (wide-sense stationary) process, $\phi$ ~ Unif$(0,2\pi)$. Suppose $X(t)$ and $\phi$ are independent random variables. I want to compute the mean and autocorrelation functions of $Y(t)$.

My attempt is (using independence of $X(t)$ and $\phi$, and $E(X(t)) = 0$)

$$ E(Y(t))= E(A X(t) \cos(\omega t + \phi)) = A \cdot E(X(t)) \cdot E(\cos(\omega t + \phi)) = 0$$

Similarly we have

$$ R_Y(t,t+\tau)= E(A X(t) \cos(\omega t + \phi) \cdot A X(t+\tau) \cos(\omega (t + \tau) + \phi)) = 0$$

Is this a correct use of independence? If not, why?

$\endgroup$
1
  • 1
    $\begingroup$ Thanks. But this also means that the autocorrelation function is zero too which I don't think is right. Am I missing something? $\endgroup$
    – XXX1010
    Mar 27 at 18:49

1 Answer 1

2
$\begingroup$

Your use of independence for computing $E[Y(t)]$ is correct. However, for the autocorrelation you get

$$\begin{align}E[Y(t)Y(t+\tau)]&=A^2E[X(t)X(t+\tau)\cos(\omega t+\phi)\cos(\omega(t+\tau)+\phi)]\\&=A^2E[X(t)X(t+\tau)]\cdot E[\cos(\omega t+\phi)\cos(\omega(t+\tau)+\phi)]\\&=A^2R_X(\tau)E[\cos(\omega t+\phi)\cos(\omega(t+\tau)+\phi)]\tag{1}\end{align}$$

Note that in general $X(t)$ and $X(t+\tau)$ are not independent, so $E[X(t)X(t+\tau)]\neq E[X(t)]E[X(t+\tau)]$.

You can compute the expectation in $(1)$ by using a trigonometric identity to write the product of the two cosines as a sum of two cosines. If you do it right, it should turn out that $E[Y(t)Y(t+\tau)]$ only depends on $\tau$ but not on $t$.

$\endgroup$
1
  • $\begingroup$ I see thank you! $\endgroup$
    – XXX1010
    Mar 27 at 21:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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