Digital Signal Processing Reference
In-Depth Information
or equivalently,
X
1
þ X
2
2
X
1
X
2
0
:
(6.47)
From here, we have:
:
X
1
þ X
2
2
X
1
X
2
(6.48)
Both left terms of (
6.48
) are the autocorrelation functions for
t ¼
0, as shown
below:
X
1
ðt
1
ÞX
1
ðt
1
þ
0
Þ¼X
1
¼ R
XX
ð
0
Þ;
X
2
ðt
2
ÞX
2
ðt
2
þ
0
Þ¼X
2
¼ R
XX
ð
0
Þ:
(6.49)
Knowing that the right-side termfrom(
6.48
) is equal to |
R
XX
(
t
)|, where
t ¼ t
2
t
1
,
and using (
6.49
), we obtain the desired result:
2
R
XX
ð
0
Þ
2
R
XX
ðtÞ
j
j:
(6.50)
P.2
An autocorrelation function is an even function,
R
XX
ðtÞ¼R
XX
ðtÞ:
(6.51)
For a WS process, we can write:
R
XX
ðtÞ¼EXðt
1
ÞXðt
2
Þ
f
g ¼ EXðt
3
ÞXðt
4
Þ
f
g;
(6.52)
where
t
1
¼ t;
t
2
¼ t þ t:
(6.53a)
t
3
¼ t;
t
4
¼ t t:
(6.53b)
Placing (
6.53b
) into (
6.52
), we get:
R
XX
ðtÞ¼EXðt
3
ÞXðt
4
Þ
f
g ¼ EXðtÞXðt tÞ
f
g ¼ R
XX
ðtÞ;
(6.54)
P.3
The value of an autocorrelation function in
t ¼
0 is a mean squared value of a
process, that is a power of the process,
R
XX
ð
0
Þ¼X
2
ðtÞ:
(6.55)
Search WWH ::
Custom Search