National University of Computer and Emerging Sciences

School of Engineering Islamabad Campus

Serial No:
EE323 Analogue & Digital Final Exam
Communication Total Time: 03 Hours
Friday, February 12, 2021 Total Marks: 100
Course Instructor
________________
Dr. Shahzad Saleem & Dr. Farhan Khalid Signature of Invigilator

_____________________________________________ _____________________
Student Name Roll No Section Signature

DO NOT OPEN THE QUESTION BOOK OR START UNTIL INSTRUCTED.

Instructions:
1. Verify at the start of the exam that you have a total of four (4) questions and a
formulae sheet printed on twenty-one (21) pages including this title page.
2. Attempt all questions on the question-book and in the given order.
3. The exam is closed books, closed notes. Please see that the area in your threshold
is free of any material classified as ‘useful in the paper’ or else there may a charge
of cheating.
4. Read the questions carefully for clarity of context and understanding of meaning
and make assumptions wherever required, for neither the invigilator will address
your queries, nor the teacher/examiner will come to the examination hall for any
assistance.
5. Fit in all your answers in the provided space. You may use extra space on the last
page if required. If you do so, clearly mark question/part number on that page to
avoid confusion.
6. Use only your own stationery and calculator. If you do not have your own
calculator, use manual calculations.
7. Use only permanent ink-pens. Only the questions attempted with permanent ink-
pens will be considered. Any part of paper done in lead pencil cannot be claimed
for rechecking.

Q-1 Q-2 Q-3 Q-4 Total

Total
20 20 25 35 100
Marks
Marks Obtained

Assessment of CLO 1 3 4 5

Vetted By: ____________________________Vetter Signature: ____________________

 National University of Computer and Emerging Sciences
School of Engineering Islamabad Campus
Q1.
CLO-01: Explain the fundamental concepts of communication systems
Answer the following short questions. [3+2+2+2+3+2+2+2+2 = 20 Marks]
a) Describe the basic principle of nonuniform quantization.
Basic principle is to use more levels at regions with large probability density
function (pdf)
i. Concentrate quantization levels in areas of largest pdf
ii. Or use fine quantization (small step size) for weak signals and coarse
quantization (large step size) for strong signals
b) Describe the function of a waveform coder?
A waveform coder or baseband binary transmitter transforms a stream of bits into a
physical waveform suitable for transmission over a channel
c) Indicate the essential blocks in a PCM system.
A PCM system comprises following essential blocks
i. Sampler
ii. Quantizer
iii. Encoder
d) Distinguish a differential PCM (DPCM) system from a conventional PCM system.
In a conventional PCM system after sampling the information signal, each sample
is quantized independently using a scalar quantizer. However, In DPCM, the
difference between two adjacent samples is quantized. Because two adjacent
samples are highly correlated, their difference has small variations. Therefore, to
achieve a certain level of performance, fewer levels (and therefore fewer bits) are
required to quantize it. This means that DPCM can achieve performance levels at
lower bit rates than PCM.
e) Describe the advantages of the vector representation of digital signal waveforms.
1. It provides a compact characterization of signal sets for transmitting digital
information over a channel, and it simplifies the analysis of their performance.
2. Using vector representation, waveform communication channels are
represented by vector channels. This reduces the complexity of analysis
considerably
f) Recognize the number of matched filters/correlators which will be used for the
detection of 4-ary PAM signals.
Regardless of the value of M, PAM is one-dimensional scheme, i.e., N = 1.
Therefore, we only need a single matched filter/correlator to demodulate it.
g) Distinguish between the dimensionality of binary antipodal and binary orthogonal

Binary antipodal signaling has a dimension = 1, whereas, binary orthogonal

signaling.

signaling has dimension = 2.

h) Explain (in words) what is meant by the maximum frequency deviation ∆ in an

Δ
FM system.

the carrier frequency, i.e., max[| − |].

represents the maximum value of the instantaneous frequency deviation from

i) How does increasing or decreasing the value of the modulation index affect the
bandwidth of an FM signal?

=2 + 1 !.
Increasing would increase the bandwidth of the FM signal and vice versa, as
indicated by the Carson’s rule, i.e.,

Final Exam Fall 2020 Page 2 of 11

 National University of Computer and Emerging Sciences
School of Engineering Islamabad Campus

Q2.
CLO-03: Analyze the performance of analog communication systems in the presence of
random noise

a) Consider a random process " = # sin 1000( , where # is a uniform random variable
[10+10 = 20 Marks]

distributed on [−1,1]. If this process is passed through an ideal bandpass filter with a
bandwidth of 1 kHz, then
i. Calculate the autocorrelation function of the output process * . (5)
ii. Calculate the powers both in the input and output processes. (5)

a) The autocorrelation function of the output process * is given by

,- . = / 01 [2- ]. 1

Now,

2- = 23 |4 |5 ,

where

1, 200 ≤ | | ≤ 1200
4 =6 .
0, otherwise

Hence,

2- = 23 ,

where we have used the fact that for an ideal filter |4 |5 = 4 = 1. Now we know
that

23 = /[,3 . ].

Therefore,

,3 + ., = >[# sin 2( × 500 + . ⋅ # sin 2( × 500 ]

#5
= >B cos 2( × 500. − cos 2( × 500 2 + . D.
2

"5
= Ecos 2( × 500. − cos 2( × 500 2 + . F ⋅ > B D,
2

where

"5 1 1 1 1
> B D = >[" 5 ] = G35 + >["]5 = G35
2 2 2 2 2

because the process " has a zero-mean. Hence,

G35
,3 + ., = Ecos 2( × 500. − cos 2( × 500 2 + . F.
2

Therefore, given process is not stationary.

By taking the average autocorrelation function of ,3 + ., because it is periodic, we get

Final Exam Fall 2020 Page 3 of 11

 National University of Computer and Emerging Sciences
1 MN 1 MN
School of Engineering Islamabad Campus
,H3 . = G35 I L cos 2( × 500. O + L cos 2( × 500 2 + . O P
JK K JK K

1
= G35 Q cos 2( × 500. + sin 2( × 500 2JK + . − cos 2( × 500. R
2000(JK

which is dependent on . only. Hence, we have

23 = / ,H3 . ,

G35
23 = QS − 500 + S + 500
2
T
+ S + 500 U V1KKKMN − S + 10W U 0V1KKKMN
2000(JK
1
− S − 500 + S + 500 R,
2000(JK

Because we have already evaluated 2- = 23 , therefore, from (1) we get

,- . = / 01 23 = ,H3 .

Power of the input process is X3 = ,H3 0 = G35 and power of the output process is
X- = ,H3 0 = G35 .
b)

b) Consider a message signal Y

2Z = 100[∏ 100[ .
which is a wide-sense stationary random process with a PSD

We also know that all the realizations of the message process satisfy max |] | = 6. We
want to transmit this message to a destination via a channel with 50-dB attenuation and
additive white noise with the power spectral density 2_ = 10015 W/Hz. We also want to
achieve an SNR at the modulator output of at least 60 dB. Calculate the required transmitter
power and channel bandwidth if we employ the following modulation schemes (4+2+4)
1. DSB AM.
2. SSB AM.
3. Conventional AM with a modulation index equal to 0.8.

Here 2Z = 100[ ∏ 100[ is nonzero for −5000 ≤ ≤ 5000, therefore, ! =

5000 Hz.

2 Xc Xc
` a = = = 10d Xc .
K! 2 × 10 × 5 × 10W
015
b

XM
Since the channel attenuation is 50 dB, it follows that
10 log = 50
Xc

Xc = 100g XM .
therefore,

2
Hence,
` a = 10W XM .
b

1. For DSB-SC AM

Final Exam Fall 2020 Page 4 of 11

 National University of Computer and Emerging Sciences
2 2
School of Engineering Islamabad Campus
` a = ` a = 10W XM ~60 = 10j .
h b

XM = 1000 Watts.
Therefore,

BW = 2! = 10[ Hn.

2. For SSB AM

2 2
` a = ` a = 10W XM ~60 = 10j .
K b

XM = 1000 Watts.
Therefore,

BW = ! = 5 kHz.

3. For conventional AM with q = 0.8

2 2
` a = s` a ,
h b

where

q5 XZt
s=
1 + q5 XZt

with

XZ XZ
XZt = = .
max|] | 5 36
gKKK gKKK
XZ = L 2Z O =L 100[ O = 1 Watt
0gKKK 0gKKK

1
XZt = .
36

0.8 × Wj
1
0.01778
5
s= = = 0.02
1 + 0.8 5 ×
1 1.01778
Wj

2
` a = 0.02 × 10W XM = 10g
h

XM = 5000 Watts

BW = 2! = 10[ Hz

Final Exam Fall 2020 Page 5 of 11

 National University of Computer and Emerging Sciences
School of Engineering Islamabad Campus
Q3. CLO-04: Perform analog-to-digital conversion of a deterministic or a random signal
with encoding

a) The source "

[15+10= 25 Marks]

2, | | < 100
is a stationary Gaussian source with zero mean and power spectral density
23 =6 .
0, otherwise

variance, the optimal output level spacing Δ = 0.5860 and distortion y = 0.03744. (10)
i. Compute an 8-level optimal quantizer for this source, where for a Gaussian source of unit

First, we compute the variance G3 in the given process.

ii. Predict the resulting distortion if the quantizer in Part i is employed? (5)
5
1KK
i.

G35 = X3 = L 2O = 400
01KK
Δ = √400 × 0.5860 = 11.72.
1
|}1 = −|}d = −3 × 11.72 − × 11.72 = −41.020
2
1
|}5 = −|}~ = −2 × 11.72 − × 11.72 = −29.30
2
1
|}W = −|}j = −1 × 11.72 − × 11.72 = −17.58
2
1
|}[ = −|}g = − × 11.72 = −5.86.
2
The resulting distortion is y = G 5 × 0.03744 = 400 × 0.03744 = 14.976.
b) In a PCM system, the signal " has a bandwidth of 5 kHz. This signal is sampled with a guard
ii.

band of bandwidth !€ =2 kHz. It is known that the sampled values have a probability density

1
function

, −2 ≤ | ≤ 0
| = • 3
3 1
, 0<|≤2
6
The quantizer uses = 128 levels. After quantization, the quantized levels are binary
encoded and transmitted.
i. Compute the resulting bit rate in bits per second? (2)
ii. Compute the resulting SQNR (in dB)? (3)
iii. Predict the required transmission bandwidth? (2)
iv. If the available transmission bandwidth is 70 kHz, predict the maximum achievable

‚ = 2! + !€ = 2 × 5000 + 2000 = 12000 samples/sec.

SQNR (in dB)? (3)

, = ƒ ‚ . = log 5 128 × 12000 = 84000 bps

i.

X3 = †05 | 5 | O| = †05 | 5 O| + †K | 5 O| = + = Watts

5 K 1 51 d [ [
ii. 3 W j ‡ ‡ W

W×[Š ×[
Therefore,
3 × 4‰ × X3
2ˆ , = = W
= 4.2144~42 dB.
|5 25
Required transmission bandwidth = 5 = 42 kHz.
c

To have a bandwidth of 70 kHz, therefore, using the equation

iii.

, ƒ ‚ ƒ 2! + !€
iv.

70000 = = = ,
2 2 2
we want to find the largest integer ƒ that satisfies above equation.
Clearly ƒ = 14 is the answer which gives !€ = 0 Hz. Since ƒ = 14 is an
increase of 7 bits/sample compared to ƒ = 7 the resulting SQNR is 42 dB
higher or 42 + 42 = 84 dB.
Final Exam Fall 2020 Page 6 of 11
 National University of Computer and Emerging Sciences
School of Engineering Islamabad Campus
Q4.
CLO-05: Explain digital modulation schemes and detection of digital signals
[14+21 = 35 Marks]

Estimate a set of orthonormal waveforms (basis functions) ‹_ , Œ = 1, … , for

a) Consider the three signal waveforms given in Figure 1 below.
i.
the signals shown in Figure 1 using the Gram-Schmidt orthogonalization procedure
(formulas are given at the end). (8)
ii. Express the three signal waveforms as vectors Ž1 , Ž5 and ŽW in the N-dimensional
signal space, using the orthonormal waveforms determined in Part a). (3)
iii. Recognize the signal space diagram (constellation diagram) showing the signal
vectors (signal points) Ž1 , Ž5 and ŽW in the N-dimensional signal space. (3)

2 0
1 2 3 t

−2
0 1 2 3 t

0
1 t

−2

Figure 1: Signal waveforms.

Final Exam Fall 2020 Page 7 of 11

 National University of Computer and Emerging Sciences
School of Engineering Islamabad Campus

Final Exam Fall 2020 Page 8 of 11

 National University of Computer and Emerging Sciences
School of Engineering Islamabad Campus
b)
i. Estimate the minimum distance O _ for binary antipodal signaling (e.g., binary
PAM, binary ASK) assuming the energy per bit, >b 1. The constellation diagram
for binary antipodal signaling is shown below. (3)
•>b •>b
Ž5 0 Ž1
Figure 2: Constellation diagram for binary antipodal signaling.

ii. Recognize the constellation diagram for binary orthogonal signaling (e.g., binary
PPM, binary FSK), clearly mentioning the coordinates of each signal point and
estimate the minimum distance O _ assuming the energy per bit, >b 1. (5)

Final Exam Fall 2020 Page 9 of 11

 National University of Computer and Emerging Sciences
School of Engineering Islamabad Campus
iii. Comparing the constellations of Part a) and Part b), infer whether binary antipodal
signaling or binary orthogonal signaling provide better performance in terms of
lower probability of error (assuming the same value of >b for both modulation
schemes). Explain briefly in terms of O _ and the effect of noise on the received
signal. (4)

Final Exam Fall 2020 Page 10 of 11

 National University of Computer and Emerging Sciences
School of Engineering Islamabad Campus
iv. Describe an advantage and a disadvantage of 4-PSK (QPSK) modulation in
comparison with 8-PSK by analyzing their constellation diagrams shown in the
Figure 3 below. Assume the signal energy >‚ to be the same for both constellations
(4)

Figure 3: 4-PSK and 8-PSK signal constellations.

1
v. Estimate the average energy > ‰ ∑Z’1‖Ž ‖5 for the 8-PAM constellation
Z
shown in the figure below (assuming equally probable signals). The location of
each signal point is mentioned in the figure below. (5)

−7d −5d −3d −d 0 d 3d 5d 7d

Figure 4: 8-PAM constellation.

Final Exam Fall 2020 Page 11 of 11