Scribd
Upload
112 views

SSP Assignment Problems - Final

This document outlines an assignment on statistical signal processing due on September 12, 2011. It includes 23 compulsory and optional questions on topics such as: 1. Analyzing a 2-channel QMF bank with block diagrams and derivations. 2. Explaining the frequency domain characteristics of downsampling with an integer factor. 3. Deriving an expression for the output autocorrelation of a filtered random process.

Uploaded by

Vadivelan Adaikkappan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
112 views

SSP Assignment Problems - Final

This document outlines an assignment on statistical signal processing due on September 12, 2011. It includes 23 compulsory and optional questions on topics such as: 1. Analyzing a 2-channel QMF bank with block diagrams and derivations. 2. Explaining the frequency domain characteristics of downsampling with an integer factor. 3. Deriving an expression for the output autocorrelation of a filtered random process.

Uploaded by

Vadivelan Adaikkappan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

ASSIGNMENT AP9114 STATISTICAL SIGNAL PROCESSING

Due Date: 09.12.2011 (By 10AM)


Compulsory Questions:
1. With a neat block diagram and relevant mathematical derivations, analyze the 2- Channel QMF
bank.
(10)
2. With neat sketches and necessary expressions, explain the frequency domain
of a down sampler with an integer factor of M.
3. Derive an expression for output autocorrelation when a random process is filtered.

characteristics
(10)
(10)

4. The power spectrum of a wide sense stationary process x(n)is


Px (ej) = [25-24cos]/[26-10cos] Determine the whitening filter H(z) that produces unit
variance white noise when the input is x(n). Also determine the impulse response of the
whitening filter.
(8)
5. Derive the variance expression and explain briefly about the method of modified periodogram
averaging and also bring out the difference between periodogram averaging method and
modified periodogram averaging methods.
(10)
6. The signal x(n) is a causal single pulse of length N with unit amplitude. Use Pronys method to
model x(n) as the unit sample response of a linear shift invariant filter having one pole and one
zero.
(7 )
7. Let x(n) be a random process with autocorrelation sequence rx(k) = (0.2)|k| . Find the reflection
coefficients for a 2nd order predictor and draw the lattice filter network.
(7 )
8. Derive the transfer function for a non-causal Wiener filter and also determine the Minimum
mean square error.
(8 )
9. Starting from the basic principles of steepest descent adaptive algorithm, derive the expression
for the error at time n, (n) interms of step size, .
(8 )
10. Derive the expression for the mean square error using LMS algorithm. Compare LMS and
Normalized LMS algorithms.
(10)
Any Five Questions:
11. Given the impulse response of the filter, h(n) = {1,-2,3,-5,7,-8,-8,7,-5,3,-2,1}. Realize the linear
phase filter by a factor of 3.
(5)
12. An FIR low pass filter with fp = 50Hz, fs = 70Hz and fo = 10KHz is to be realized using a two
stage decimator and interpolator with an integer factor of 10. The stop band ripple and pass band
ripple are given as 0.001 and 0.01 respectively. Determine the overall complexity of the filter?
Perform the same calculation if the integer factor is reduced to 6 and comment on the FOPS. (5)

13. What is meant by a stationary random process? Differentiate between SSS and WSS random
process and mention its properties.
(5)
14. Using the Yule Walker equation, determine the coefficient values for an AR (1) process. (5)
15. Determine the mean and autocorrelation of the signal x(n) = Ae j(n0 + ) where A and 0 are
fixed constants and is a random variable that is uniformly distributed over the interval - to
,i.e., the probability density function for is
f () = (2)-1 ; -
0
; otherwise
(5)
16. With necessary derivation, explain briefly the periodogram smoothing using Blackman- Tukey
method.
17. The input to a LSI filter with unit sample response h(n) = (n) + 0.5(n-1) + 0.25(n-2) is a zero
mean wide sense stationary process with autocorrelation rx(k) = (0.5) |k|
(i)
What is the variance of the output process?
(ii)
Find the autocorrelation of the output process, ry(k) for all k.
(5)
18. Consider a first order AR process that is generated by the difference equation
y(n) = ay(n-1) + w(n) where |a|<1 and w(n) is a zero mean white noise random process with
variance w2. Determine the
(i)
Unit sample response of the filter that generates y(n) from w(n)
(ii)
Autocorrelation of y(n)
(iii)
Power spectrum of y(n).
(5)
19. Given rx(0) = 1 and the first three reflection coefficients are 1 = 2 = 0.5 and 3 = 0.25. Find
the corresponding autocorrelation sequence.
(6)
20. Suppose that the first five values in the autocorrelation sequence for the process x(n) are rx = [3,
9/4, 9/8, 9/16, 9/32,.]T. Use modified Yule Walker equation method to find ARMA(1,1)
model for x(n).
(5)
21. Derive the normal equation and the minimum error for all pole model using Pronys method (5)
22. Suppose that a signal d(n) is corrupted by noise, x(n) = d(n) + w(n) where r w(k)=0.5(k) and
rdw(k) = 0. The signal d(n) is an AR(1) process that satisfies the difference equation d(n) =
0.5d(n-1) + v(n) , where v(n) is white noise with variance v2= 1. Design a 1st order FIR linear
predictor W(z) = w(0) + w(1)z-1 for d(n) and also find the minimum mean square prediction
error = E{[d(n+2) d(n+2)]2}. Assume that w(n) and v(n) are uncorrelated.
(6)
23. Consider the system shown in the figure below for estimating a process d(n) from x(n)
(Refer Exercise Problem in Optimum Filter chapter)
If d2= 4 and rx = [1, 0.5, 0.25]T ; rdx =[-1,1]T. Find the value of a(1) that minimizes the mean
square error = E{|e(n)|2} and also find the minimum mean square error.
(6)

You might also like