www.binils.

com
Anna University | Polytechnic | Schools
*X85082* Reg. No. :

Question Paper Code : x 85082

M.E./M.Tech. DEgREE ExaMiNaTioNs, NoVEMBER/DECEMBER 2020
First semester
applied Electronics
aP 5152 – aDVaNCED DigiTal sigNal PRoCEssiNg
(Common to M.E. Communication systems/M.E. Communication
and Networking/M.E. Digital signal Processing/M.E. Electronics and
Communication Engineering /M.E. Electronics and Communication
Engineering (Industry Integrated))

(Regulations 2017)
Time : Three Hours Maximum : 100 Marks

nswer all questions

A
Pa – (10�2=20 Marks)
rt
A
1. tate the Weiner-Khitchine relation and discuss about it.
S
2. ist the necessary and sufficient condition for a random process is said to be
L
wide sense stationary process.

3. Define bias, unbiased and asymptotically unbiased estimate.

4. Differentiate Non-parametric and Parametric method of power spectrum

estimation.

5. Mention any two applications where Kalman filter is used.

6. What is lattice filter structure ? What is the advantage of such structure ?

7. What is the advantage of adaptive filters over optimum filters ?

8. Bring out the limitations of steepest descent algorithm.

9. Mention the necessity for multistage implementation of sampling rate conversion.

10. ist out few applications of multirate signal processing.

L
 www.binils.com
Anna University | Polytechnic | Schools
85082 -2- *X85082*
X
Pa –B (5�13=65 Marks)

rt
11. a) i) With necessary equations, explain in detail about special types of
random process. (7)
ii) Derive the spectral factorization of the power spectrum Px(ejw). (6)
( R)

O
b) i) x(t) is a wide sense stationary process with autocorrelation function
sin(2000πt ) + sin(1000πt ) .
R x( τ ) = 10

2000πt
The process x(t) is sampled at rate 1/Ts = 4,000 Hz, yielding the discrete-

time-process xn. What is the autocorrelation function R [k] of xn ? (5)

X
ii) tate and explain Parseval’s theorem with its properties. (8)
S
12. a) Explain the Bartlett and Welch method of smoothing the periodogram and
evaluate the performance measures. (13)
( R)
O
b) Explain the periodogram method of spectrum estimation in detail and also
obtain the variance of the periodogram. (13)

13. a) i) Explain about stochastic RM model in detail. (7)

A
A
ii) How the yule walker method to solve the R model parameters ? (6)
A
( R)
O
b) Explain the Kalman filter estimation approach in detail. Derive the
expression for Kalman gain that minimizes mean square error. (13)

14. a) tarting from the basic principles, derive the M weight update equation
S
L
S
and explain the M adaptive algorithm in detail. (13)
L
S
( R)
O
b) Explain in detail about exponentially weighted R and sliding window
LS
R algorithms. (13)
LS
15. a) With neat diagram and relevant expressions, explain the time domain and
frequency domain characteristics of a decimater with a decimation factor
of ‘D’. (13)
( R)
O
b) Draw the polyphase decomposed structure of a 15-tap F R filter with a
I
decomposition factor of ‘3’. (13)
 www.binils.com
Anna University | Polytechnic | Schools
*X85082* -3- 85082

X
Pa –C (1�15=15 Marks)

rt
16. a) ssume that v(n) is a real-valued zero-mean white aussian noise with
A
G
sv2 = 1, x(n) and y(n) are generated by the equations
x(n) = 0.5x(n – 1) + v(n),

y(n) = x(n – 1) + x(n).

i) Compute the power spectrum of x(n) and y(n). (7)
ii) Compute Ry(k); for k = 0, 1, 2, 3. (4)
iii) RM (1, 1) spectral estimate. (4)
A
A
( R)
O
b) simple averaging filter is defined as
A
1
y[n ] =
N
(x[n − 1] + ... + x[n − N ])

This is clearly an F R filter.

I
i) et N = 4. Determine the transfer function, its zeros and poles ; (4)
L
ii) determine a general form for zeros and poles for any N ; (4)
iii) By comparing y(n) and y[n – 1] determine a recursive implementation.
lso the transfer function, together with its zeros and poles of the recursive
A
implementation. ooking at this example, can we say that “any” recursive
L
filter is R ? (7)
II
–––––––––––––