Line 46: Line 46:
 
----
 
----
  
Problem 2. [50 pts] <br>
+
Problem 2. [40 pts] <br>
Consider a finite-length sinewave of the form below where <math> k_{o} </math> is an interger in the range <math> 0 \leq k_{o} \leq N-1 </math>.<br/>
+
(a) Let <math>x[n]</math> and <math>y[n]</math> be real-valued sequences both of which are even-symmetric: <math>x[n]=x[-n]</\math> and <math>y[n]=y[-n]</\math>. Under these conditions, prove that  
<center><math> x[n] = e^{j 2 \pi \frac{k_{o}}{N} n} \{ u[n] - u[n-N] \} </math>  (2)</center><br/>
+
<math>r_{xy}[l]=r_{yx}[l]</\math> for all <math>l</\math>.<br>
In addition, h[n] is a causal FIR filter of length L, where L < N. In this problem <math> y[n]=x[n] \star h[n] </math> is the linear convolution of the causal sinewave of length N in Equation (1) with a causal FIR filter of length L, where L < N.<br/>
+
(b) Express the autocorrelation sequence r_{zz}[l] for the complex-valued signal <math>z[n]=x[n]+jy[n]</\math> where <math>x[n]</\math> and <math>y[n]</\math> are real-valued sequences, in terms of <math>r_{xx}[l],r_{xy}[l],r_yx[l]</\math> and <math>r_{yy}[l]</\math>.<br>
<center><math> y[n]=x[n] \star h[n] </math></center><br/>
+
(a) The region <math> 0 \leq n \leq L-1 </math> corresponds to partial overlap. The covolution sum can be written as:<br/>
+
<center><math> y[n]=\sum_{k=??}^{??} h[k] x[n-k] \ \ partial \ \ overlap:\ 0 \leq n \leq L-1  </math> (3) </center><br/>
+
Determine the upper and lower limits in the convolution sum above for <math> 0 \leq n \leq L-1 </math><br/>
+
(b) The region <math> L \leq n \leq N-1 </math> corresponds to full overlap. The convolution sum is: <br/>
+
<center><math> y[n]=\sum_{k=??}^{??} h[k]x[n-k] \ \ full \ \ overlap: \ L \leq n \leq N-1 </math>  (4)</center><br/>
+
(i) Determine the upper and lower limits in the convolution sum for <math> L \leq n \leq N-1 </math>.<br/>
+
(ii) Substituting x[n] in Eqn (1), show that for this range y[n] simplifies to:<br/>
+
<center><math> y[n]=H_{N}(k_{o}) e^{j 2 \pi \frac{k_{o}}{N} n} \ \ for L \leq n \leq N-1  </math>(5)</center>  <br/>
+
where <math> H_{N}(k) </math> is the N-point DFT of h[n] evaluated at <math> k = k_{o} </math>. To get the points, you must show all work and explain all details. <br/>
+
(c) The region <math> N \leq n \leq N+L-2 </math> corresponds to partial overlap. The convolution sum: <br/>
+
<center><math> y[n] = \sum_{k=??}^{??} h[k]x[n-k] \ \ partial \ \ overlap: \ N \leq n \leq N+L-2 </math>(6)</center>  <br/>
+
Determine the upper and lower limits in the convolution sum for <math> N \leq n \leq N+L-2 </math>. <br/>
+
(d) Add the two regions of partial overlap at the beginning and end to form: <br/>
+
<center><math> z[n] =y[n]+y[n+N]=\sum_{k=??}^{??}h[n]x[n-k] \ \ for: \ 0 \leq n \leq L-1 </math>(7)</center> <br/>
+
(i) Determine the upper and lower limits in the convolution sum above. <br/>
+
(ii) Substituting x[n] in Eqn (1), show that for this range z[n] simplifies to: <br/>
+
<center> <math> z[n] = y[n]+y[n+N]=H_{N}(k_{o})e^{j2 \pi \frac{k_{o}}{N} n} \ \ for \ 0 \leq n \leq L-1 </math> (8)</center><br/>
+
where <math> H_{N}(k) </math> is the N-point DFT of h[n] evaluated at <math> k = k_{o}</math> as defined previously. <br/>
+
(e) <math> y_{N}[n] </math> is formed by computing <math> X_{N}(k) </math> as an N-pt DFT of x[n] in Enq (2), <math> H_{N}(k) </math> as an N-pt DFT of h[n], and then <math> y_{N}[n] </math> as the N-pt inverse DFT of <math> Y_{N}(k) = X_{N}(K)H_{N}(k) </math>. Write a simple, closed-form expression for <math> y_{N}(k) </math>. Is <math> z[n]=y_{N}[n] + y[n+N] \ \ for \ 0 \leq n \leq N-1 </math>? <br/>
+
 
+
  
 +
:'''Click [[ECE_PhD_QE_CNSIP_2015_Problem1.1|here]] to view student [[ECE_PhD_QE_CNSIP_2015_Problem1.1|answers and discussions]]'''
 
----
 
----
  
  
 
[[ECE_PhD_Qualifying_Exams|Back to ECE QE page]]
 
[[ECE_PhD_Qualifying_Exams|Back to ECE QE page]]

Revision as of 23:58, 17 February 2019


ECE Ph.D. Qualifying Exam

Communicates & Signal Process (CS)

Question 2: Signal Processing

August 2011




Problem 1. [60 pts]
In the system below, the two analysis filters, $ h_0[n] $ and $ h_1[n] $, and the two synthesis filters, $ f_0[n] $ and $ f_1[n] $,form a Quadrature Mirror Filter (QMF). Specially,
$ h_0[n]=\dfrac{2\beta cos[(1+\beta)\pi(n+5)/2]}{\pi[1-4\beta^2(n+5)^2]}+\dfrac{sin[(1-\beta)\pi(n+0.5)/2]}{\pi[(n+.5)-4\beta^2(n+.5)^3]},-\infty<n<\infty with \beta=0.5 $
$ h_1[n]=(-1)^n h_0[n] $ $ f_0[n]=h_0[n] $ $ f_1[n]=-h_1[n] $
The DTFT of the halfband filter $ h_0[n] $ above may be expressed as follows:
$ H_0(\omega)= \begin{cases} e^{j\dfrac{\omega}{2}} |\omega|<\dfrac{\pi}{4},\\ e^{j\dfrac{\omega}{2}} cos[(|\omega|-\dfrac{\pi}{4})], \dfrac{\pi}{4}<|\omega|<\dfrac{3\pi}{4} \\ 0 \dfrac{3\pi}{4}<|\omega|<\pi \end{cases} $
Wan82_ECE538_problem1.PNG Consider the following input signal
$ x[n]=16\dfrac{sin(\dfrac{3\pi}{8}n)}{\pi n}\dfrac{sin(\dfrac{\pi}{8}n)}{\pi n}cos(\dfrac{\pi}{2}n) $
HINT: The solution to problem is greatly simplified if you exploit the fact that the DTFT of the input signal $ x[n] $ is such that $ X(\omega)=X(\omega-\pi) $.
(a) Plot the magnitude of the DTFT of $ x[n] $, $ X(\omega) $, over $ -\pi<\omega<\pi $. Show all work.
(b) Plot the magnitude of the DTFT of $ x_0[n] $, $ X_0(\omega) $, over $ -\pi<\omega<\pi $. Show all work.
(c) Plot the magnitude of the DTFT of $ x_1[n] $, $ X_1(\omega) $, over $ -\pi<\omega<\pi $. Show all work.
(d) Plot the magnitude of the DTFT of $ y_0[n] $, $ Y_0(\omega) $, over $ -\pi<\omega<\pi $. Show all work.
(e) Plot the magnitude of the DTFT of $ y_1[n] $, $ Y_1(\omega) $, over $ -\pi<\omega<\pi $. Show all work.
(f) Plot the magnitude of the DTFT of the final output $ y[n][n] $, $ Y(\omega) $, over $ -\pi<\omega<\pi $. Show all work.

Click here to view student answers and discussions

Problem 2. [40 pts]
(a) Let $ x[n] $ and $ y[n] $ be real-valued sequences both of which are even-symmetric: $ x[n]=x[-n]</\math> and <math>y[n]=y[-n]</\math>. Under these conditions, prove that <math>r_{xy}[l]=r_{yx}[l]</\math> for all <math>l</\math>.<br> (b) Express the autocorrelation sequence r_{zz}[l] for the complex-valued signal <math>z[n]=x[n]+jy[n]</\math> where <math>x[n]</\math> and <math>y[n]</\math> are real-valued sequences, in terms of <math>r_{xx}[l],r_{xy}[l],r_yx[l]</\math> and <math>r_{yy}[l]</\math>.<br> :'''Click [[ECE_PhD_QE_CNSIP_2015_Problem1.1|here]] to view student [[ECE_PhD_QE_CNSIP_2015_Problem1.1|answers and discussions]]''' ---- [[ECE_PhD_Qualifying_Exams|Back to ECE QE page]] $

Alumni Liaison

Basic linear algebra uncovers and clarifies very important geometry and algebra.

Dr. Paul Garrett