Assignment 1
Assignment 1
Assignment 1
The last date for assignment submission to avoid negative marking is 31st October. After that,
the following formula will be used to calculate your marks:
𝑅𝑀 = 𝑂𝑀 ∗ exp(−0.05 ∗ 𝑑)
where 𝑅𝑀 is the revised marks, 𝑂𝑀 is the obtained marks, and d is the number of days after
31st October.
For example, if you submit your assignment on or before 31st October and get 60 marks in the
assignment, we will consider 60 marks only. However, if you submit on 1st November, your
revised marks will be 60 ∗ 𝑒 −0.05∗1 = 57.07. Similarly, if you submit on 2nd November, your
revised marks will be 60 ∗ 𝑒 −0.05∗2 = 54.29 and so on.
Therefore, it would be worth considering the assignment submission well before 31st October.
𝑁0
where w(t) is AWGN with zero mean and P.S.D and 𝑠(𝑡) = 𝑢(𝑡) − 2𝑢(𝑡 − 1) + 𝑢(𝑡 −
2
2) where 𝑢(𝑡) represents the unit step function.
The received signal is passed through a filter, and the output is sampled to yield a decision
statistic. An ML decision rule is employed on the decision statistic. The setup is shown in
figure.1.
𝐸
(a) For ℎ(𝑡) = 𝑠(−𝑡), derive an expression for error probability as a function of 𝑁𝑏 if 𝑇𝑜 =
𝑜
1.
(b) Find the best and worst possible choices of sampling time in the interval 0 ≤ 𝑇𝑜 ≤ 2.
(Ignore the practical issues involved in the realization of an anti-causal filter)
Figure .1
2. Consider a baseband 4-ary system using 4 discrete amplitude levels. The receiver model is
as shown in figure.2, the operation of which is governed by the following assumptions:
(a) The signal component in the received wave is
𝑡
𝑚(𝑡) = ∑ 𝑎𝑛 sinc ( − 𝑛)
𝑇𝑠
𝑛
where 𝑇𝑠 is the symbol duration.
(b) The amplitude levels are 𝑎𝑛 = ± 𝐴⁄2, ± 3𝐴⁄2
(c) The 4 levels are equiprobable, and the symbols transmitted in adjacent time slots are
statistically independent.
(d) The channel noise 𝑤(𝑡) is white and Gaussian with zero mean and power spectral
density 𝑁𝑜 ⁄2.
(e) The low-pass filter is ideal with bandwidth 𝐵𝑇 = 1⁄2𝑇𝑠 .
(f) The threshold levels used in the decision device are 0, ±𝐴.
Figure .2
Show that the average probability of symbol error in this system is given by:
3 𝐴
𝑃𝑒 = 𝑄( )
2 2𝜎
where 𝜎 is the standard deviation of the noise at the input of the decision device.
where 𝑎𝑘 ∈ {−1, +1} with equal probability, and 𝑔(𝑡) is a rectangular pulse of amplitude
𝐴 and width 𝑇𝑏 . The received signal is given by
∞
where 𝑤(𝑡) is zero mean AWGN process with P.S.D. 𝑁0 /2. The response of matched filter
is given by ℎ(𝑡) = 𝑔(𝑇𝑏 − 𝑡). Suppose the clock that specifies the sampling of the
matched output is offset from the optimum sampling time by 10 percent.
(a) Show that the output of the matched filter at the time instant 𝑡 = (𝑚 − 0.1)𝑇𝑏 is
9 1
𝑦𝑚 = 10 𝑎𝑚−1 𝐴2 𝑇𝑏 + 10 𝑎𝑚−2 𝐴2 𝑇𝑏 + 𝑛𝑚 .
4. In line-of-sight microwave radio transmission, the signal arrives at the receiver via two
propagation paths: the direct path and a delayed path that occurs due to signal reflection
from surrounding terrain. Suppose that the received signal has the form
where 𝑠(𝑡) = ∑∞ 𝑘=−∞ 𝑎𝑘 𝑔(𝑡 − 𝑘𝑇𝑠 ) , 𝑎𝑘 ∈ {−1, +1} with equal probability and 𝑔(𝑡) is a
rectangular pulse of duration 𝑇𝑠 and energy 𝐸𝑠 , 𝛽 is the attenuation (𝛽 < 1) of the
𝑁0
secondary path, and 𝑤(𝑡) is AWGN process with zero mean and P.S.D. . If the matched
2
filter response ℎ(𝑡) = 𝑔(𝑇𝑠 − 𝑡)
(a) Show that the output of the matched filter at the time instant 𝑡 = 𝑚𝑇𝑠 is
𝑦𝑚 = 𝑎𝑚−1 𝐸𝑠 + 𝛽𝑎𝑚−2 𝐸𝑠 + 𝑛𝑚
(b) Determine the probability of error for 𝑎𝑚−1 in terms of 𝐸𝑠 /𝑁0 assuming threshold, 𝜆 =
0.