411 2004 MidtermSolutions

________________________________________________________________
Instructor: K. T. Wong
Time allowed: 2 hours.

Closed-book closed-notes examination.
A formula sheet is provided; no other formula sheet is allowed.
Please answer all questions.
The maximum points possible are 40.
No electronic devices or calculators allowed.

_________________________________________________________________

Since X is uniformly distributed in [-3.5, 3.5],
| |
( ) ( ) ( ) ( )
( )
( )
( )
( )
( ) . of ues other val all for , 0
7
1
7
1
) (
3 , 2 , 1 , 0 , 1 , 2 , 3 , 5 . 0 5 . 0
otherwise if , 0
5 . 3 , 5 . 3 if ,
7
1
) (
5 . 0
5 . 0
5 . 0
5 . 0
y y Y P
dx dx x f
k k x k P k Y P
x
x f
k
k
k
k
X
X
= =
=
= =
= + < = =

=

+

+

(8 points)
Consider a 7-level uniform mid-tread quantizer with input-output relation
illustrated in the diagram below. (This diagram is not drawn to scale.) A random
variable, uniformly distributed between [-3.5, 3.5], is applied as the quantizer input.
Find the probability density function of the random variable at the quantizer output.
Simplify your answer as much as possible for full credit. You MUST show the
STEP-BY-STEP FULL details of the derivation for credit.
2
7
y, quantizer
output
x, quantizer input 0
2
5
2
3
2
7
2
5
2
3
2
2

2
3

2
3

(8 points).
A music signal has a maximum frequency of 20 KHz
of 3 volts. This music signal is applied to a delta-modulator
per second. Discuss the choice of an appropriate step-size
avoid slope-overload distortion.
and a maximum amplitude
with bit rate at 100 kilobits
for the delta-modulator to
e
dt
t dx
T
s
) (
max
. The
omponent of the input x(t)
(t)/dt = -2f
m
A
m
sin(2f
m
t).
{(20000) (3)} volts/second.
lts to avoid slope-overload
(8 points). A PCM system uses a uniform quantizer followed by an 8-bit
binary encoder. The bit rate of the system equals 40 megabits per second. What is the
maximum message bandwidth for which the system operates satisfactorily?
To avoid slope-overload distortion, we must choos
maximum value of dx(t)/dt occurs at the highest frequency c
and maximum amplitude. Let x(t) = A
m
cos(2f
m
t), then dx
Correspondingly, max |dx(t)/dt| = 2 max(f
m
A
m
) = 2 max
Hence, choose a step-size
( )

2 . 1
000 , 100
60000 2
= vo
distortion.

Let the message bandwidth be W. Then, sampling the message signal at its Nyquist rate
and using an n-bit code to represent each sample of the message signal, we find that the
bit duration equals T
b
= T
s
/ n = 1 /(2nW). The bit rate is R
b
= 1/T
b
= 2nW. The maximum
value of the message bandwidth is thus W
max
= 40,000,000 / 2 / 8 = 2,500,000 (Hz).

(12 points). An uncorrelated random discrete-time sequence of logical values
{A
n
} takes on the values 0 or 1 with equal probability. A precoder converts {A
n
} into a
discrete-time sequence of voltage levels {B
n
}, using the polar (antipodal) scheme:
where c and d are positive real-value numbers. Derive, step-by-step, the autocorrelation
function R
B
(k) of {B
n
}.

( )
| |
( )
2
2 2
2 2
2 2
2 2
2 2
2
2
4
1
) ( ) ( ) ( ) ( ) ( ) (
) ( ) ( ) ( ) ( ) (
) , ( ) ( ) , ( ) (
) , ( ) ( ) , ( ) (
, 0 For
2
1
) ( ) ( ) ( ) 0 (
, 0 For
|
.
|
\
|
= + =
= = + = = +
= = + = = =
= = + = = +
= = + = = = =
+ =
= = + = = =
=

d c
cd d c
c B P d B cP d d B P c B P d c
d B P d B P d c B P c B P c
c B d B cP d d B c B P d c
d B d B P d c B c B P c B B E k R
k
d c
B d B P d B c B P c R
k
k n n k n n
k n n k n n
k n n k n n
k n n k n n k n n B
k n n k n n B

=
=
=
0 for ,
1 for ,
n
n
n
a d
a c
b
(4 points).
The raised-cosine pulse, defined in the time-domain and frequency-domain as
follows, is parameterized by [0,1].

A larger value of would smooth the raised-cosine spectrums transition from the
passband to the stopband, but at the cost of a larger bandwidth. The corresponding time-
domain effect is lower sidelobes in between the time-sampling grid-points {n T
b
, n}, for
a larger value of . (Note that time-sampling will lose all time-domain information in the
original continuous-time signal between sampling instants {n T
b
, n}.)
Give one advantage obtained in trade-off against the aforementioned disadvantage
of a larger bandwidth at a larger value of . Your explanation should be in full details for
full credit. An under-developed cryptic reply (say, consisting of just a few catchwords)
may receive no credit.

The advantage is enhanced robustness against non-ideal time-sampling, when the
actual sampling instants deviate from the ideal sampling instants {n T
b
, n}. A smaller
sidelobe will then reduce the ISI. Such deviation may occur due to: (1) imperfect
circuitry in pulse-generator and/or time-sampler and/or other modules (for example to
lower hardware cost), (2) Doppler effects unknown to the receiver.

411 2004 MidtermSolutions

Uploaded by

Copyright:

Available Formats

411 2004 MidtermSolutions

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

411 2004 MidtermSolutions

Uploaded by

Copyright:

Available Formats

________________________________________________________________

You might also like