Expected Value of the Rayleigh Random Variable

Sahand Rabbani
We consider the Rayleigh density function, that is, the probability density function of the Rayleigh random
variable, given by
f
R
(r) =
r

2
e

r
2
2
2
Note that this is radial, so we consider f
R
(r) for r > 0. We endeavor to nd the expectation of this random
variable. This is given by the integral
E{R} =
_

0
rf
R
(r)dr =
_

0
r
2

2
e

r
2
2
2
dr
I oer three solutions.
Solution with Integration by Parts
First, we will consider an integration by parts approach. Recall
_
u dv = uv
_
v du
Careful choice of u and v gives
E{R} =
1

2
_

0
r
..
u
re

r
2
2
2
dr
. .
dv
That is
u = r
du = dr
v =
2
e

r
2
2
2
dv = re

r
2
2
2
dr
This gives
E{R} =
1

2
_

2
re

r
2
2
2

0
+
_

0

2
e

r
2
2
2
dr
_
= re

r
2
2
2

0
. .

+
_

0
e

r
2
2
2
dr
. .

The expectation is the sum of the two terms and . Let us consider :
= lim
r
_
re

r
2
2
2
_
lim
r0
_
re

r
2
2
2
_
The rst limit is indeterminant, but we can see that the exponential term decays to zero faster than the
polynomial term grows. Formally, we use LH opitals rule to evaluate the limit:
lim
r
_
re

r
2
2
2
_
= lim
r
r
e
r
2
2
2
= lim
r
(d/dr)r
(d/dr)e
r
2
2
2
= lim
r
1
(r/
2
)e
r
2
2
2
= 0
1
S. Rabbani Expected Value of the Rayleigh Random Variable
The second term of the limit can be evaluated by simple substitution:
lim
r0
_
re

r
2
2
2
_
= re

r
2
2
2

r=0
= 0
Thus,
= 0 0 = 0
Our problem reduces to,
E{R} =
_

0
e

r
2
2
2
dr =
This integral is known and can be easily calculated. By symmetry, it is clear that,
=
_

0
e

r
2
2
2
dr =
1
2
_

r
2
2
2
dr
Consider now

2
=
_
1
2
_

r
2
1
2
2
dr
1
__
1
2
_

r
2
2
2
2
dr
2
_
=
1
4
_

r
2
1
+r
2
2
2
2
dr
1
dr
2
We now convert to polar coordinates. Specically,
2
= r
2
1
+ r
2
2
. By the Jacobian, we have that dr
1
dr
2
=
d d. This gives

2
=
1
4
_
2
0
_

0
e

2
2
2
d d
=
1
4
__
2
0
d
___

0
e

2
2
2
d
_
=
1
4
2
_

0
e

2
2
2
d
=

2
_

2
e

2
2
2
_

0
=

2
[0 (
2
)] =

2

2
Thus,
=
_

2
=
_

2
And, nally,
E{R} =
_

2
Solution with Fourier Transforms
We note from symmetry that
E{R} =
_

0
r
2

2
e

r
2
2
2
dr =
1
2
_

r
2

2
e

r
2
2
2
dr
It is now evident that
E{R} =
1
2
_

_
r
2

2
e

r
2
2
2
_
e
2(0)r
dr = F
_
r
2
2
2
e

r
2
2
2
_
(s)

s=0
2
S. Rabbani Expected Value of the Rayleigh Random Variable
We note that
F
_
1
2
2
e

r
2
2
2
_
(s) =
1

2
e
2
2

2
s
2
By the derivative theorem, we have that
F
_
r
2
2
2
e

r
2
2
2
_
=
_
j
2
_
2
d
2
ds
F
_
1
2
2
e

r
2
2
2
_
(s)
Dierentiating twice gives
F
_
r
2
2
2
e

r
2
2
2
_
(s) =
_
j
2
_
2
_
1

2
_
(4
2

2
)(1 4
2

2
s
2
)e
2
2

2
s
2
Evaluating this expression at s = 0 gives
E{R} = F
_
r
2
2
2
e

r
2
2
2
_
(s)

s=0
=
_
j
2
_
2
_
1

2
_
(4
2

2
)(1 4
2

2
s
2
)e
2
2

2
s
2

s=0
=
_

2
Solution with the Gamma Function
From before, we have,
E{R} =
_

0
r
2

2
e

r
2
2
2
dr
We now make a variable substitution given by
s =
r
2
2
2
r =

2s
ds =
r dr

2
The integral becomes
E{R} =

2
_

0

s e
s
ds
We recognize this as the Gamma Function (x)
1
evaluated at x =
3
2
:
E{R} =

2
_

0
s
3
2
1
e
s
ds =

2
_
3
2
_
It is well known that
(x + 1) = x (x)
and that

_
1
2
_
=

1
The Gamma Function is dened as
(x) =
Z

0
t
x1
e
t
dt
3
S. Rabbani Expected Value of the Rayleigh Random Variable
Thus

_
3
2
_
=
1
2

_
1
2
_
=

2
and, nally
E{R} =
_

2
4