Name:

Student Number:

Page 1

EE456: Digital Communications

MIDTERM EXAM, 5:30PM7:30PM, October 26, 2011
(2 hours, open-textbook)
Examiner: Ha H. Nguyen
Note: There are 3 questions. All questions are of equal value (with part marks
indicated) but not necessarily of equal difficulty. Full marks shall only be given
to solutions that are properly explained and justified.
1. The input noise x(t) applied to the filter below is modeled as a wide-sense stationary,
zero-mean white, Gaussian random process with power spectral density N0 /2. Let y(t)
denote the random process at the output of the filter.

A linear filter
H( f )
1

y (t )

x (t )

Figure 1: A filter considered in Question 1.

[3]

(a) Find and sketch the power spectral density of the random process y(t).

[3]

(b) What are the average DC level and the average power of the noise y(t)?

[2]

(c) Find the

Z autocorrelation function of the random process y(t).
Hint:
x2 cos xdx = 2x cos x + (x2 2) sin x.

[2]

1
(d) Now suppose that the output noise is sampled every Ts = 2W
seconds to obtain
the noise samples y(kTs ), k = 0, 1, 2, . . .. Determine whether the noise samples
are statistically independent. Explain your answer.

EE456Digital Communications

Electrical & Computer Engineering

Name:

Student Number:

Page 2

2. Consider the signal space diagram in Figure 2, where the waveforms of the two basis
functions are also shown. The signal set is used for data transmission over an AWGN
channel in which the noise is zero-mean and has two-sided power spectral density of
N0 /2. As usual, s1 (t) is used for the transmission of bit 0 and s2 (t) is for the
transmission of bit 1. Furthermore, the two bits are equally likely and the bit
duration is Tb .

2 (t )

1
Tb

Tb

1
Tb

s2 (t ) 1T

2V Tb

2V Tb

1 (t )

4V Tb

1
Tb

2V Tb

s1 (t ) 0T

Tb

Figure 2: Signal space diagram considered in Question 2.

[2]

(a) Determine and neatly sketch the two signals s1 (t) and s2 (t). Without performing
the integration but by observing Figure 2, can you tell whether s1 (t) and s2 (t)
are orthogonal ? Why?

[2]

(b) Clearly draw the optimum decision boundary and label the two decision regions.
Write the expression for the optimum decision rule in terms of r1 and r2 , the

EE456Digital Communications

Electrical & Computer Engineering

Name:

Student Number:

Page 3

projections of the received signal on 1 (t) and 2 (t). Simplify the decision rule as
much as you can.
[2]

(c) Let N0 = 1012 watts/Hz. What is the lowest voltage level V you can set in order
to transmit at the rate of 1 Mbps (megabits per second) and with a probability
of error P [error] 105 . Hint: Q1 (105 ) 4.25.

[2]

(d) Clearly explain how you can perform axis rotation to obtain new basis functions
1 (t) and 2 (t) so that the optimum receiver can be implemented with only one
matched filter. Draw the block diagram of such an optimum receiver. Determine
and plot the impulse response of the matched filter. Note: You do not need to
determine the exact value of the threshold for the comparison block.

[2]

(e) Your friend argues that, instead of performing axis rotation, she can move signal
s2 (t) in the signal space diagram in such a way that the new signal, called sM
2 (t),
has the same component with s1 (t) along 2 (t). Assuming that the energy of
s2 (t) cannot be changed, clearly identify the location of sM
2 (t). What would you
tell your friend about the performance (over an AWGN channel) of her modified
signal set as compared to the performance of the original signal set? Explain your
answer.
3. (AMI-RZ ) Alternate-Mark-Invert Return-to-Zero is a binary line coding scheme that
has memory. The output signal is determined from the sources bit stream as follows:
If the bit to be transmitted is a 0, then the signal is 0 volts over the bit period of
Tb seconds.
If the bit to be transmitted is a 1, then the signal is either +V volts or V volts
over the first half of Tb and it is zero over the second half of Tb . Over the first
half of Tb , the signal is +V volts if V was used previously to represent bit 1 and
it is V volts if +V was previously used to represent bit 1. Hence the name and
mnemonic for the modulation.

[3]

(a) Draw the AMI-RZ waveform for the information sequence {11010010}. Assume
that the amplitude level before the transmission of the above sequence is V
volts. Comment on the synchronization property of AMI-RZ scheme.

[5]

(b) Identify and clearly plot the three signals used in AMI-RZ scheme. Show a signal
space representation of these three waveforms and associate bits 0 and 1 with the
corresponding signals. Then clearly draw the minimum-distance boundary and
the block diagram of the corresponding bit-by-bit minimum-distance receiver. If
you have a matched filter in your block diagram, clearly show the impulse response
of the filter.

[2]

(c) Assume that the AMI-RZ modulation is used over an AWGN channel with twosided PSD of N0 /2 (watts/Hz). Determine the conditional bit error probability
P [error|0T ]. Remark: The error expression should be in the form of a Q-function,
whose argument depends on V , Tb and N0 .

EE456Digital Communications

Electrical & Computer Engineering

Name:

s1 (t )

s2 ( t )

A
Student Number:

t
POTENTIALLY
FACTS:
0 USEFUL
T 2

T 2

Page 4

Fourier transform:
Z

j2f t

x(t) =

X(f )e

FT

df X(f ) =

x(t)ej2f t dt

(1)

t = Tb
r 0 0D
r (t )
Minimum-distance
h (t )decision
= s1 (Tb rule
t ) for equally likely bits:
r < 0 1D
1D

(r1 s11 )2 + (r2

(r1 s21 )2 + (r2 s22 )2 .
2 (st12) )2 R
1
0

(2)

2 (t ) 2

1T s 2 (t )

0
1

s2 (t )
d2
1

2
r(t )

( s21 , s22 )
0

( r1 , r2 )

s1 (t )

1 (t ) 1

( s11 , s12 )
s1 (t ) 0T

d1

1 (t )

0 2

Rotation of axes:

2 (t )
s2 ( t )

1 (t )

2 (t )
s1 (t )

1 (t )

1 (t)
2 (t)

cos sin
sin cos

1 (t)
2 (t)

(3)

1
f m (m),

volts
1
EE456Digital Communications

Electrical & Computer Engineering

m, (volts)