# Jointly WSS

1. Jan 13, 2013

### SeriousNoob

1. The problem statement, all variables and given/known data

ϕ(t) = R(t) = autocorrelation
ϕ(f) = S(f) = power spectral density
Inputs v[t] and w[t] are zero mean, jointly WSS processes with auto-correlation se-
quences ϕvv(τ ) and ϕww(τ ), cross-correlation sequence ϕvw(τ ), power spectrums ϕvv(ω)
and ϕww(ω), and cross power spectrum ϕvw(ω). Let x(t) be the output of a LTI system
h1(t) with input v(t), i.e. x(t) = h1(t)v(t), and let y(t) be the output of a LTI system
h2(t) with input w(t), i.e. y(t) = h2(t)  w(t). Assume all quantities to be complex.

(a) Show that x(t) and y(t) are jointly WSS. Determine an expression for ϕxy(τ ) and
ϕxy(ω) in terms of the input statistics and the system parameters.

(b) Is the cross power spectrum always non-negative, i.e. is ϕxy(ω) ≥ 0, for all ω? Justify

(c) Specialize this formula to obtain an expression for ϕxx(ω).

2. Relevant equations
ϕxy(τ) = E[x(t)y(t+τ)]
ϕx(f) = ϕv(f)|H|²
ϕx(f) = Fourier Transform( ϕx(t) )

3. The attempt at a solution
I'm at a complete loss. For jointly WSS, I need to prove x(t) and y(t) are WSS which they are because of v(t) and w(t) going through an LTI system. But I also need to show ϕxy(τ) only depends on τ.
I feel that I'm missing a crucial point and don't know how to start.

2. Jan 19, 2013

### hikaru1221

I suppose that it is trivial to show X(t) and Y(t) are WSS So the problem is only to prove X(t) and Y(t) are jointly WSS.
To do this, expand $R_{XY}(t,t+\tau)$:

$X(t) = \int_{-\infty}^{+\infty}V(t-\alpha)h_1(\alpha)d\alpha$

$Y(t+\tau) = \int_{-\infty}^{+\infty}W(t+\tau-\beta)h_2(\beta)d\beta$

Hence:

$R_{XY}(t,t+\tau) = \int_{\alpha=-\infty}^{+\infty} \int_{\beta=-\infty}^{+\infty} E[V(t-\alpha)W(t+\tau-\beta)]h_1(\alpha)h_2(\beta)d\beta d\alpha$

It is easy now :)

Last edited: Jan 19, 2013