Digital Signal Processing Reference
In-Depth Information
ð
1
e
joa
ob
e
job
e
job
2
j
e
joa
ob
e
joðbyþaÞ
1
f
X
ðoÞ¼
2
d
y ¼
¼
sin
ðboÞ:
(2.402)
1
Now we will demonstrate how the characteristic function is also useful to
express the density function of the transformed random variable
Y ¼ g
(
X
)in
terms of the PDF of the variable
X.
From
y ¼ gðxÞ;
(2.403)
we get:
x ¼ g
1
ðyÞ:
(2.404)
Considering (
2.404
), we can write:
f
X
ðxÞj
x¼g
1
ðyÞ
¼ f
X
ðyÞ;
d
x ¼
g
1
ðyÞ
d
y;
(2.405)
Placing (
2.405
) into (
2.386
), we arrive at:
1
1
f
Y
ðoÞ¼Ef
e
joY
e
jogðxÞ
f
X
ðyÞg
1
ðyÞ
d
y ¼
e
joy
hðyÞ
d
y:
g¼
(2.406)
1
1
Comparing the two integrals in (
2.406
), we can express density function as:
f
Y
ðyÞ¼hðyÞ¼f
x
ðyÞg
1
ðyÞ:
(2.407)
The difficulty arises when the transform is not unique, because it is not possible
to simply express the transformed PDF in the form (
2.407
).
Example 2.8.10
Find the density function from Fig.
2.52
, using (
2.401
).
Solution
ð
ð
1
aþb
1
b
y a
b
f
Y
ðoÞ¼Ef
e
joY
e
jogðxÞ
f
X
ðxÞ
d
x ¼
e
joy
f
X
g¼
d
y
1
ab
ð
ð
aþb
aþb
e
joy
1
2
1
b
e
joy
hðyÞ
d
y:
¼
d
y ¼
(2.408)
ab
ab
Search WWH ::
Custom Search