Joint CDF of correlated samples derived from a gaussian process

From: Gaurav Chandra (gchandra_at_ti.com)
Date: 08/23/04 

Date: Mon, 23 Aug 2004 12:43:22 +0000 (UTC)

Hi,

I want to solve the following problem, would be grateful if someone
can help:

I have a white gaussian discrete time process (sequence of i.i.d.
gaussian random samples). I pass this through a finite impulse
response filter. That is, if input is x_n (nth sample) and output
is y_n, and filter has (m+1) taps,
 
y_n = some linear combination of x_(n-m) through x_n

Output samples are now correlated.

Q: Can I somehow derive the Joint CDF of any two samples taken from
this output process, given that the input pdf is gaussian?

Does this CDF (and PDF) satisfy some known properties (e.g. jointly
gaussian?).

Any help would be greatly appreciated.

Gaurav

Quantcast