1
Reg. No. :
Question Paper Code :50546
B.E./B.Tech. DEGREE EXAMINATIONS, APRILMAY 2024.
iFifth /Sixth Semester
Electronics and Communiation Engineering
CEC 332ADVANCED DIGITAL SIGNAL PROCESSING
(Common to : Electronics and Telecommunication Engineering)
(Regulations 2021)
Time: Three hours Maximum: 100 marks
Answer ALL questions.
PART A (10 x 2=20 marks)
1. Consider a discrete time signal x2]=2, 1, 2, 2] given as input to an LTI
system with impulse response of hr]= 1,2, 1, 2]. Determine the output of the
system.
2. Compute the Z-transform of the signal xn]= a'un].
3. What are the conditions, to be. satisfied for a random process to be wide sense
stationary?.
4. What is Ergodic random process? Give an grgnple.
5. Draw the lattice filter implementation of FIRWeper flter.
6. Mention the advantages of Kalman filter when compared to Wiener flter?
7. Draw the block diagram of adaptive channel equalization technique.
8. Represent the weight vector úpdate equation of LMS algorithm and the
importance of step size in weight vector adjustment.
9 Mention the differences between parametric and non paranetric method of
SgU POwer.8pectrum estimation.
est
10. What is ARMA power spctrum estimation method?
PART B (6 × 13 = 65 marks)
downsampler and illustrate the need
11. (a) Derive the frequency spectrum of a
for an anti-aliasing filter in a decimator.
Or
coding. Derive the
Explain the significance of filter banks in sub band
(b) quadrature mirror two
flter impulse response of perfect reconstruction
channelGlter bank.
the random process is
12. (a) For each of the following, determine whether or not
Wide sense stationary and Mean ergodic. 1
density
) xn)= A, wheré A is a random: variable with probability
(6)
fünction A(a)..
(i) ln)= Acosna, +¢) where is a random variable that is
(7)
uniformly distributed between-r and .
Or
(b) MA
Consider an MA(O-process that is generated by the difference equation,
where n) is zero mean white noiso with variance
i) Find the unit sample respons¿ of the-Glter that generates yn) from
(4)
ala).
(ii) Find the autocorrolation of y(u).: (4)
(i) Find the power spectrum of sr). (5)
13. (a) Let dn) be an AR(1) process with an autocorrelation sequence,
rk)a. Here, 0<a<l, and dn) is observed in the presence of
uncorrelated white noise, un), that has a variance of o. Design a ürst
order FIR Wioner filter and derive he filter function for a =0.8 and
e e l . rove that the filter is a low pass ilter by finding the
frequency
response of the filter.
Or
50546
according to the difference
(6) Consider an AR(p) proce8s that is generated
equation,
x(a)=So(k)e(n -k)+an)
equation and -computation
Derive the state, equation, observation the given AR(P) process,
equations for the discrete Kalman filter for
the condition for convergence
14. (a) Explain LMS algorithm in detail and derive
of the LMS algorithm.
Or
the time update cquation for the coefficient vector W, in
(b) Derive
exponentially weighted RLS algorithm.
xn]. Prove that the periodogram is
15. (a) Consider any discrete time sequence
proportional to the squared magnitude of
the DTET of the sequence. Also
discuss briefly on the performance of the periodogram.
Or
estimation and specify the spectrum
(b) What is Auto Regressive spectrum various AR methods. available to
equation of the same. Discuss the
estimate the power spectråm.
PART C-(1x 15 = 15 marks)
shift-invariant systemn having a system
16. (a) Suppose we are given ainear
function,
correlated noise xn with an
that is excited by zero mean exponenrially
autocorrelation sequence,
Let sn) be the output process, y(n)x<n)" h(r)
(4)
) Find the power pectrum,e), of ylu) 2
(4)
(Gi) Find the autocorrelation,seqyence r, ()., of y"
and y() (4)
(üii) Find the cross-correlation, rk, between x(r)
which ia the
(iv) Pind the cro85-power"'spectral density, Pz)
(3)
tránsforn of the (k).
50546
Design an efficient sampling rate convértor to increase the samping rate
by a factor The input discrete signal to the convertor is
=3, 8, 3, 3,3, 3) and filter coefficients of anti-aliasing flter is
hn=(1 1; 1) and the flter coefficients of anti-imaging filter is
gn=(3, 2,1). Compute the output. of the designed sampling rate
convertor.
Pade pz Ap
Dr. V MOHAN
sUPERVIsOR REF.No: 3340068
ASSOCIATE PROFESSOR,
Engineer1n
Communlcation
and
of Electronics
Erngineering
Oept College of
- 620 012.
saranathari
Trychirappalli
Dr. V. MOHAN
SUPERVISOR REF.No: 3340068
ASSOCIATE PROFESSOR,
Communication EngineA(
deot of Elettronles and
Jarunathan College of Erngineerins
tiHychitappalli-620 012
Dr, V. MOHAN
SUPERVISOR REFNÍ: 3340068
J ASSOCIATE PROFESSOR,
Jept of Electronics and Comtnunication Engin á/h: 50546
Anathan College of ngineert
tWotehitappalli -620 012.
