EE160 Fall 2004 San Jose State University

Problem Set # 2
I. Gaussian random variable
I.1. The noise voltage in an electric circuit can be modeled as a Gaussian random variable with
mean equal to zero and variance equal to 10
8
.
(a) What is the probability that the noise exceeds 10
4
? What is the probability that it exceeds
4 10
4
? What is the probability that the noise value is between 2 10
4
and 10
4
?
(b) Given that the noise voltage is positive, what is the probability that it exceeds 10
4
?
(c) This noise passes through a half-wave rectier with characteristics
g(x) =
_
x, x > 0;
0, x 0.
Find the PDF of the rectied noise by rst nding its CDF. Why can we not use the general
expression in Eq. (4.1.10) of the textbook?
(d) Find the mean of the rectied noise in the previous part.
(e) The noise passes now through a full-wave rectier dened by g(x) = |x|. Find the density
function of the rectied noise and its mean value.
Solution:
(a) The random variable X is Gaussian with zero mean and variance
2
= 10
8
. Thus
p(X > x) = Q
_
x

_
,
and
p(X > 10
4
) = Q
_
10
4
10
4
_
= Q(1) = .159
p(X > 4 10
4
) = Q
_
4 10
4
10
4
_
= Q(4) = 3.17 10
5
p(2 10
4
< X 10
4
) = 1 Q(1) Q(2) = .8182
(b)
p(X > 10
4

X > 0) =
p(X > 10
4
, X > 0)
p(X > 0)
=
p(X > 10
4
)
p(X > 0)
=
.159
.5
= .318
(c) y = g(x) = xu(x). Clearly f
Y
(y) = 0 and F
Y
(y) = 0 for y < 0. If y > 0, then the equation
y = xu(x) has a unique solution x
1
= y. Hence, F
Y
(y) = F
X
(y) and f
Y
(y) = f
X
(y) for y > 0.
F
Y
(y) is discontinuous at y = 0 and the jump of the discontinuity equals F
X
(0).
F
Y
(0
+
) F
Y
(0

) = F
X
(0) =
1
2
In summary the PDF f
Y
(y) equals
f
Y
(y) = f
X
(y)u(y) +
1
2
(y)
The general expression for nding f
Y
(y) can not be used because g(x) is constant for some
interval so that there is an uncountable number of solutions for x in this interval.
(d)
E[Y ] =
_

yf
Y
(y)dy
=
_

y
_
f
X
(y)u(y) +
1
2
(y)
_
dy
=
1

2
2
_

0
ye

y
2
2
2
dy =

2
(e) y = g(x) = |x|. For a given y > 0 there are two solutions to the equation y = g(x) = |x|, that
is x
1,2
= y. Hence for y > 0
f
Y
(y) =
f
X
(x
1
)
|sgn(x
1
)|
+
f
X
(x
2
)
|sgn(x
2
)|
= f
X
(y) +f
X
(y)
=
2

2
2
e

y
2
2
2
For y < 0 there are no solutions to the equation y = |x| and f
Y
(y) = 0.
E[Y ] =
2

2
2
_

0
ye

y
2
2
2
dy =
2

2
I.2. X is an N(0,
2
) random variable. This random variable is passed through a system whose
input-output relation is given by y = g(x). Find the PDF or the PMF of the output random variable
Y that results in the following cases:
(a) Square-law device: g(x) = ax
2
, where a is a constant.
(b) Hard limiter:
g(x) =
_

_
a, x > 0;
0, x = 0;
b, x < 0,
where a and b are constants.
Solution:
(a) y = g(x) = ax
2
. Assume without loss of generality that a > 0. Then, if y < 0 the equation
y = ax
2
has no real solutions and f
Y
(y) = 0. If y > 0 there are two solutions to the system,
namely x
1,2
=
_
y/a. Hence,
f
Y
(y) =
f
X
(x
1
)
|g

(x
1
)|
+
f
X
(x
2
)
|g

(x
2
)|
=
f
X
(
_
y/a)
2a
_
y/a
+
f
X
(
_
y/a)
2a
_
y/a
=
1

ay

2
2
e

y
2a
2
(b) In the case of the hard limiter
p(Y = b) = p(X < 0) = F
X
(0) =
1
2
p(Y = a) = p(X > 0) = 1 F
X
(0) =
1
2
Thus F
Y
(y) is a staircase function and
f
Y
(y) = F
X
(0)(y b) + (1 F
X
(0))(y a) =
1
2
[(y b) +(y a)] .
I.3. (Spring 2003) A binary random variable S takes values S = +1 and S = 1 with equal
probability. Let R = S +N denote the random variable that is observed after passing S through a
noisy channel. The channel acts as adding a zero-mean Gaussian random variable N with variance

2
= 1/2. Variables S and N are assumed independent.
N
Noisy
channel
S R
(a) Find and sketch carefully the probability density function of R. (Hint: The conditional PDF
of R, given a value of S, f
R
(r|S = s), is that of a linear transformation of N. Then use the
total probability theorem.)
(b) To estimate the value of S, the random variable R is passed through a threshold device that
produces the estimate

S = sgn(R), where
sgn(x)

=
_
+1, x 0;
1, x < 0.
Find the probability of error, P() = Pr{

S = S}. (Hint: Use conditional probability and the

total probability theorem. An error is made when R has opposite sign from S.)
Solution:
(a)
f
R
(r) =
1
2
f
R
(r|S = +1) +
1
2
f
R
(r|S = 1)
=
1
2

e
(r1)
2
+
1
2

e
(r+1)
2
f
R
(r)
r
-1 +1
(b)
P() =
1
2
Pr{R > 0|S = 1} +
1
2
Pr{R 0|S = +1}
= Q
_
0 + 1
_
1/2
_
= Q(

2) Q(1.4) = 7.865 10
2
I.4. A zero-mean white Gaussian noise process with power spectral density
N
0
2
passes through
an ideal lowpass lter of bandwidth B.
(a) Find the autocorrelation function of the output process Y (t)
(b) Assume that =
1
2B
. Determine the PDF of the random variables Y (t) and Y (t + ). Are
they independent? (Hint: Use the autocorrelation function of Y (t).)
Solution:
(a) S
X
(f) =
N
0
2
, R
X
() =
N
0
2
(). The autocorrelation function and the power spectral density of
the output are given by
R
Y
(t) = R
X
() h() h(), S
Y
(f) = S
X
(f)|H(f)|
2
With H(f) = (
f
2B
) we have |H(f)|
2
=
2
(
f
2B
) = (
f
2B
) so that
S
Y
(f) =
N
0
2
(
f
2B
)
Taking the inverse Fourier transform of the previous we obtain the autocorrelation function of the
output
R
Y
() = 2B
N
0
2
sinc(2B) = BN
0
sinc(2B)
(b) The output random process Y (t) is a zero mean Gaussian process with variance

2
Y (t)
= E[Y
2
(t)] = E[Y
2
(t +)] = R
Y
(0) = BN
0
The correlation coecient of the jointly Gaussian processes Y (t +), Y (t) is

Y (t+)Y (t)
=
COV (Y (t +)Y (t))

Y (t+)

Y (t)
=
E[Y (t +)Y (t)]
BN
0
=
R
Y
()
BN
0
With =
1
2B
, we have R
Y
(
1
2B
) = sinc(1) = 0 so that
Y (t+)Y (t)
= 0. Hence the joint probability
density function of Y (t) and Y (t +) is
f
Y (t+)Y (t)
=
1
2BN
0
e

Y
2
(t+)+Y
2
(t)
2BN
0
Since the samples are Gaussian and uncorrelated they are also independent.
I.5. Additive white Gaussian noise X(t), with zero mean and power spectral density S
X
(f) = N
0
/2,
is applied to a linear lter whose impulse response h(t) is shown below. A sample Y is taken of the
output of the lter at time T.
t
h(t)
0 T
1/T
(a) Find the mean
Y
and the variance
2
Y
of Y , in terms of N
0
and T.
(b) Determine an expression of the probability density function of Y .
Solution
(a) Since
X
= 0, it follows that
Y
= 0. Also, for a sample of a zero-mean random process,

2
Y
= E{Y
2
(T)} = R
Y
(0) =
_

S
Y
(f)df
=
N
0
2
_

|H(f)|
2
df =
N
0
2
_

|h(t)|
2
dt =
N
0
2
_
1
T
_
2
T =
N
0
2T
.
(b) The input of the linear lter is Gaussian noise. It follows that the output is also Gaussian
noise. As shown in part (a), the sample Y = Y (T) has mean
Y
= 0 and variance
2
Y
=
N
0
2T
.
Therefore
f
Y
(y) =
1
_
N
0
T
e

T
N
0
y
2
.
II. Stationary random processes
II.1. X(t) is a stationary process with power spectral density S
X
(f). This process passes through
the system shown in the gure below:
X(t)
Y(t)
Delay
T
d
dt
(a) Is Y (t) stationary? Why?
(b) What is the power spectral density of Y (t)?
(c) What frequency components cannot be present in the output process and why?
Solution:
(a) The impulse response of the system is h(t) = L[(t)] =

(t) +

(t T). It is a LTI system so

that the output process is a stationary. This is true since Y (t +c) = L[X(t +c)] for all c, so if
X(t) and X(t +c) have the same statistical properties, so do the processes Y (t) and Y (t +c).
(b) S
Y
(f) = S
X
(f)|H(f)|
2
. But, H(f) = j2f +j2fe
j2fT
so that
S
Y
(f) = S
X
(f)4
2
f
2

1 +e
j2fT

2
= S
X
(f)4
2
f
2
[(1 + cos(2fT))
2
+ sin
2
(2fT)]
= S
X
(f)8
2
f
2
(1 + cos(2fT))
(c) The frequencies for which |H(f)|
2
= 0 will not be present at the output. These frequencies
are f = 0, for which f
2
= 0 and f =
1
2T
+
k
T
, k Z, for which cos(2fT) = 1.
II.2. A zero-mean white Gaussian noise, n
w
(t), with power spectral density N
0
/2, is passed thorugh
an ideal lter whose passband is 3-11 KHz. The output process is denoted by n(t).
(a) If f
0
= 7 KHz, nd S
nc
(f) and S
nc
(f), the power spectral densities of the in-phase and
quadrature components (these are denoted n
I
(t) and n
Q
(t) in our class notes.)
(b) Repeat part (a) with f
0
= 6 KHz.
Solution:
(a) The power spectral density of the in-phase and quadrature components is given by
S
nc
(f) = S
ns
(f) =
_
S
n
(f f
0
) +S
n
(f +f
0
) |f| < 7
0 otherwise
If the passband of the ideal lter extends from 3 to 11 KHz, then f
0
=7 KHz is the mid-band
frequency so that
S
nc
(f) = S
ns
(f) =
_
N
0
|f| < 7
0 otherwise
(b) With f
0
=6 KHz
S
nc
(f) = S
ns
(f) =
_
_
_
N
0
2
3 < |f| < 5
N
0
|f| < 3
0 otherwise
II.3. A noise process n(t) has a power spectral density given by
S
n
(f) =
_
10
8
_
1
|f|
10
8
_
, |f| 10
8
;
0, |f| 10
8
.
This noise is passed through an ideal bandpass lter with a bandwidth of 2 MHz centered at 50
MHz.
(a) Find the power content of the output process.
(b) Write the output process in terms of the in-phase and quadrature components and nd the
power content of each component. Assume that f
0
= 50 MHz.
(c) Find the power spectral density of the in-phase and quadrature components.
(d) Assume now that the lter is not ideal but described by
|H(f)|
2
=
_
|f| 49 10
6
, 49 10
6
< |f| 51 10
6
;
0, otherwise.
Repeat parts (a)-(c) under this assumption.
Solution:
(a) The power spectral density S
n
(f) is depicted in the following gure. The output bandpass
process has non-zero power content for frequencies in the band 4910
6
|f| 5110
6
. The
power content is
P =
_
4910
6
5110
6
10
8
_
1 +
f
10
8
_
df +
_
5110
6
4910
6
10
8
_
1
f
10
8
_
df
= 10
8
x

4910
6
5110
6
+ 10
16
1
2
x
2

4910
6
5110
6
+ 10
8
x

5110
6
4910
6
10
16
1
2
x
2

5110
6
4910
6
= 2 10
2
= 20 mW
!
!
!
!
!
!
!
!
!
a
a
a
a
a
a
a
a
a
510
7
510
7
10
8
10
8
(b) The output process N(t) can be written as
N(t) = N
c
(t) cos(250 10
6
t) N
s
(t) sin(250 10
6
t)
where N
c
(t) and N
s
(t) are the in-phase and quadrature components respectively, given by
N
c
(t) = N(t) cos(250 10
6
t) +

N(t) sin(250 10
6
t)
N
s
(t) =

N(t) cos(250 10
6
t) N(t) sin(250 10
6
t)
The power content of the in-phase component is given by
E[|N
c
(t)|
2
] = E[|N(t)|
2
] cos
2
(250 10
6
t) +E[|

N(t)|
2
] sin
2
(250 10
6
t)
= E[|N(t)|
2
] = 2 10
2
= 20 mW
where we have used the fact that E[|N(t)|
2
] = E[|

N(t)|
2
]. Similarly we nd that E[|N
s
(t)|
2
] =
2 10
2
= 20 mW.
(c) The power spectral density of N
c
(t) and N
s
(t) is
S
Nc
(f) = S
Ns
(f) =
_
S
N
(f 50 10
6
) +S
N
(f + 50 10
6
) |f| 50 10
6
0 otherwise
S
Nc
(f) is depicted in the next gure. The power content of S
Nc
(f) can now be found easily as
P
Nc
= P
Ns
=
_
10
6
10
6
10
8
df = 2 10
2
= 20 mW
10
8
10
6
10
6
(d) The power spectral density of the output is given by
S
Y
(f) = S
X
(f)|H(f)|
2
= (|f| 49 10
6
)(10
8
10
16
|f|) for 49 10
6
|f| 51 10
6
Hence, the power content of the output is
P
Y
=
_
4910
6
5110
6
(f 49 10
6
)(10
8
+ 10
16
f)df
+
_
5110
6
4910
6
(f 49 10
6
)(10
8
10
16
f)df
= 2 10
4

4
3
10
2
The power spectral density of the in-phase and quadrature components of the output process
is given by
S
Yc
(f) = S
Ys
(f) =
_
(f + 50 10
6
) 49 10
6
_ _
10
8
10
16
(f + 50 10
6
)
_
+
_
(f 50 10
6
) 49 10
6
_ _
10
8
+ 10
16
(f 50 10
6
)
_
= 2 10
16
f
2
+ 10
2
for |f| 10
6
and zero otherwise. The power content of the in-phase and quadrature component
is
P
Yc
= P
Ys
=
_
10
6
10
6
(2 10
16
f
2
+ 10
2
)df
= 2 10
16
1
3
f
3

10
6
10
6
+ 10
2
f

10
6
10
6
= 2 10
4

4
3
10
2
= P
Y