Digital Signal Processing Reference
In-Depth Information
If the mean value of the process is zero,
XðtÞ ¼
0, then
ðtÞ ¼ s
2
R
XX
ð
0
Þ ¼ X
2
X
;
(6.56)
where
s
X
is the variance of the process.
For
t ¼
0, from (
6.33
), we can write,
ðtÞ ¼ s
X
:
R
XX
ð
0
Þ ¼ XðtÞXðt þ
0
Þ ¼ X
2
(6.57)
P.4
If the mean value of the process is not zero, then the autocorrelation function
has a constant term which is equal to the squared mean value:
2
R
XX
ðtÞ ¼ R
X
0
X
0
ðtÞþXðtÞ
(6.58)
where
X
0
(
t
) is a zero-mean process obtained from the process
X
(
t
) by subtracting the
mean value
c
.
Consider the WS process
X
(
t
), where a mean value is a constant
c
,
E
{
X
(
t
)}
¼ c
,
We can write:
XðtÞ ¼ c þ X
0
ðtÞ;
(6.59)
The autocorrelation function of the process
X
(
t
) is
g ¼ E c þ X
0
ðtÞ
c þ X
0
ðt þ tÞ
R
XX
ðtÞ ¼ E XðtÞXðt þ tÞ
f
f
½
½
g
¼ E c
2
þ X
0
ðtÞX
0
ðt þ tÞþcX
0
ðtÞþcX
0
ðt þ tÞ
¼ c
2
þ E X
0
ðtÞX
0
ðt þ tÞ
g ¼ c
2
f
þ R
X
0
X
0
ðtÞ
:
(6.60)
P.5
If a random process is a periodic process, then its autocorrelation function also
has a periodic component of the same period as the process itself.
(a) Consider a periodic process
X
(
t
)
¼ A
cos(
o
0
t
+
y
) where
A
, and
o
0
¼
2
p
/
T
,
where
T
is a period, are constants, and
y
is a uniform random variable in the
range [0, 2
p
].
The autocorrelation function is:
R
XX
ðtÞ ¼
X
ð
t
Þ
X
ð
t
þ
t
Þ ¼
A
cos
ð
o
0
t
þ
y
Þ
A
cos
o
0
ð
t
þ
t
Þþ
y
ð
Þ
Þ
:
(6.61)
A
2
2
A
2
2
¼
cos
ðo
0
to
0
to
0
tÞþ
cos
o
0
ð
2
t þ tÞþy
ð
We can easily find that the second term is equal to zero, resulting in:
A
2
2
R
XX
ðtÞ ¼
cos
o
0
t:
(6.62)
Search WWH ::
Custom Search