Digital Signal Processing Reference
In-Depth Information
Using the characteristics of the delta function from (
2.72
), we arrive at:
"
#
2
e
jo
0
t
1
þðo
0
TÞ
2
e
jo
0
t
1
þðo
0
TÞ
2
e
jo
0
t
þ
e
jo
0
t
2
1
2
ðo
0
TÞ
2
þ
ð
o
0
T
Þ
ðo
0
TÞ
R
YY
ðtÞ¼
¼
;
2
2
1
þðo
0
TÞ
2
ðo
0
TÞ
¼
2
cos
ðo
0
tÞ:
ð
7
:
172
Þ
1
þðo
0
TÞ
The autocorrelation function (
7.171
) does not have a constant
term and
consequently
EYðtfg¼
0
:
(7.173)
The variance is:
2
ðtÞ
EYðtfg¼ EY
2
ðtÞ
¼ R
YY
ð
0
Þ¼
ð
o
0
T
Þ
s
YY
¼ EY
2
2
:
(7.174)
1
þðo
0
TÞ
The obtained results (
7.173
) and (
7.174
) are equal to the results (
7.168
) and
(
7.170
), respectively.
Exercise E.7.11
If the input process
X
(
t
) from Exercise E.7.9 has a Gaussian PDF
what will be the PDF of the output process
Y
(
t
)?
Answer
Since the input process is subjected to a LTI filtering, it follows that the
input Gaussian process is subjected to a linear transformation. As a consequence,
the output process is also a Gaussian process (see Sect.
4.3
).
The process
Y
(
t
) is a WS stationary (with a constant mean and autocorrelation
function that depends only on
t
) and, consequently, its PDF is independent of time.
This means that the PDF of the output process is:
; s
YY
Þ:
f
Y
ðyÞ¼Nð
0
(7.175)
where the variance is given in (
7.174
).
Exercise E.7.12
Consider that the input to the RC filter shown in Fig.
7.27
is a
white noise with the PSD
S
XX
(
o
)
¼ N
0
/2. Find the PSD and the autocorrelation
function of the process at the output of the filter.
Fig. 7.27
RC filter in
Exercise E.7.12
Search WWH ::
Custom Search