Digital Signal Processing Reference
In-Depth Information
or equivalently
1
2
EfX
2
R
XX
ðt; t þ tÞ¼
g
cos
o
0
t þ y o
0
ðt þ tÞy
½
ð
Þ þ
cos
o
0
t þ y þ o
0
ðt þ tÞþy
ð
Þ
1
2
EfX
2
¼
g
cos
ðo
0
tÞþ
cos
ð
2
o
0
t þ
2
y þ o
0
tÞ
½
1
2
EfX
2
¼
g
cos
ðo
0
tÞþ
cos 2
ðo
0
t þ yÞþo
0
t
½
ð
Þ
:
(6.188)
The received result indicates that the autocorrelation function depends not only
on
t
but also on
t
. Therefore, the second condition is not satisfied and the process is
not a WS stationary.
Exercise 6.8
Verify and explain whether or not each of the following can be an
autocorrelation function.
2
t
1
þ t
2
:
gðtÞ¼
(a)
(6.189)
2
t
2
1
þ t
2
:
(b)
gðtÞ¼
(6.190)
(c)
gðtÞ¼
sin
o
0
t;
o
0
¼
constant
:
(6.191)
(d)
gðtÞ¼
cos
o
0
t;
o
0
¼
constant
:
(6.192)
Answer
An autocorrelation function which is a function of a time difference
t
must
have the properties listed in Sect.
6.5.4
.
(a) This function cannot be an autocorrelation function because it does not have the
property (
6.45
)
j
gðtÞ
j gð
0
Þ:
(6.193)
> g
(0).
Additionally, this function does not have the property (
6.51
) (i.e., it is not an
even function).
(b) This function satisfies the property (
6.51
), i.e., it is an even function. However,
it cannot be an autocorrelation function because it does not have the property
(
6.45
). For example,
g
(0)
¼
0,
g
(1)
¼
1, and
g
(1)
For example,
g
(0)
¼
0,
g
(2)
¼
4/5; and thus
g
(2)
> g
(0).
(c) This function cannot be an autocorrelation function because it is not an even
function.
Search WWH ::
Custom Search