Digital Signal Processing Reference
In-Depth Information
7.4.2 Multiplication of Random Process with a Sinusoidal Signal
There are many practical applications where one may need to multiply a random
process with a sinusoidal function, as shown in Fig.
7.16
.
We can describe a time and spectral characteristics of the output process
Y
(
t
)
YðtÞ¼XðtÞU
0
cos
o
0
t:
(7.103)
The autocorrelation function is:
R
YY
ðt; t þ tÞ¼EYðtÞYðt þ tf g
¼EXðtÞU
0
cos
o
0
t
½
Xðt þ tÞU
0
cos
o
0
ðt þ tÞ
½
g
;
f
¼ U
0
cos
o
0
t
cos
o
0
ðt þ tÞEXðtÞXðt þ tÞ
f
g:
(7.104)
Using the definition for the autocorrelation function (
6.32
) and the trigonometric
U
0
R
YY
ðt; t þ tÞ¼
2
R
XX
ðt; t þ tÞ
cos
o
0
t þ
cos
ð
2
o
0
t þ o
0
tÞ
½
:
(7.105)
If the process
X
(
t
) is at least WS stationary, then
R
XX
ðt; t þ tÞ¼R
XX
ðtÞ;
(7.106)
and (
7.105
) reduces to:
U
0
2
R
XX
ðtÞ
cos
o
0
t þ
cos
ð
2
o
0
t þ o
0
tÞ
R
YY
ðt; t þ tÞ¼
½
:
(7.107)
Note that the autocorrelation function (
7.107
) is a function of time and it is
necessary to find the time average for it:
T=
ð
2
1
T
AR
YY
ðt; t þ tÞ
f
g ¼
lim
T!1
R
YY
ðt; t þ tÞ
d
t;
T=
2
T=
2
ð
U
0
U
0
1
T
¼
2
R
XX
ðtÞ
cos
o
0
t þ
2
R
XX
ðtÞ
lim
cos
ð
2
o
0
t þ o
0
tÞ
d
t;
T!1
T=
2
U
0
2
R
XX
ðtÞ
cos
o
0
t;
¼ R
YY
ðtÞ:
¼
ð
7
:
108
Þ
Search WWH ::
Custom Search