Differences

This shows you the differences between two versions of the page.

--- faq [2012/10/08 14:37] – asif
+++ faq [2012/10/25 14:47] (current) – asif
@@ Line 26: / Line 26: @@
 **Assignment 2**\\
 [[faq#QE1. For the Additional Problem, I'm unsure how to check if x[k] is WSS or not.  For 1., we only know A is a random variable with some unknown pdf fA(a).  I can't find the mean using the formula for the first moment. How should this problem be tackled?|QE1. For the Additional Problem, I'm unsure how to check if x[k] is WSS or not.  For 1., we only know A is a random variable with some unknown pdf fA(a).  I can't find the mean using the formula for the first moment. How should this problem be tackled?]]\\
+**Assignment 3**\\
+[[faq#QF1. For the additional problem, I'm thinking of how to approach part a).  Since s1(t) and  s2(t) are sinusoidal signals and are bandpass signals, can we  convert the signals to baseband by using only the magnitudes?  For s1(t), it would be sqrt(2E/T), and s2(t) would be -sqrt(2E/T)?]]\\
+[[faq#QF2. For the additional problem, I'm thinking of how to approach part a).  For the sinusoidal signals, do we assume the value is 0 at the end of period T?
+When I am trying to find the output z1(t), my result contains the sine function, but, which would be 0 if the end of period T is always at 0, like how we usually draw them in class.]]\\
+[[faq#QF3. For problem 3.6, we are given the signal waveforms s1(t) and s2(t), but not a1(t) or a2(t).  Can we assume s1(t) = a1(t) and s2(t) = a2(t)?  I was thinking we could do this because the channel effect will be filtered out by the time we sample z(t) and it is passed through the ML detector.  Is this assumption correct?]]\\
 ----
@@ Line 76: / Line 85: @@
 ----
 == QE1. For the Additional Problem, I'm unsure how to check if x[k] is WSS or not.  For 1., we only know A is a random variable with some unknown pdf fA(a).  I can't find the mean using the formula for the first moment. How should this problem be tackled?==
-E1. You do not have to find the values. Just check if there dependence on time. For (1),
+E1. You do not have to find the values. Just check there dependence on time. For the first part, for example,:\\
+. Calculate the expected value of A. It is time independent.\\
+. Calculate the expected value of x[k1] x[k2]. It is given by E{A^2} and is time independent (a special case of E{x[k1] x[k2]} = E{x[k1] x[k1 + m]} with m = 0.\\
+Thus x[k] is a WSS process.
+----
+== QF1.  For the additional problem, I'm thinking of how to approach part a).  Since s1(t) and  s2(t) are sinusoidal signals and are bandpass signals, can we  convert the signals to baseband by using only the magnitudes?  For s1(t), it would be sqrt(2E/T), and s2(t) would be -sqrt(2E/T)?==
+F1. The circuit is given in Fig. 2. You do not need to any  down-frequency conversion. For part (a), the impulse response of the  matched filter is given by h1(t) = s1(T - t) and h2(t) = s2(T - t).
+== QF2.  For the additional problem, I'm thinking of how to approach part a).  For the sinusoidal signals, do we assume the value is 0 at the end of period T? When I am trying to find the output z1(t), my result contains the sine function, but, which would be 0 if the end of period T is always at 0, like how we usually draw them in class.==
+F2. Yes. You should assume that both s1(t) and s2(t) are zero outside the duration T.
-. Calculate the expected value of A. It is time independent.
+== QF3.  For problem 3.6, we are given the signal waveforms s1(t) and s2(t), but not a1(t) or a2(t).  Can we assume s1(t) = a1(t) and s2(t) = a2(t)?  I was thinking we could do this because the channel effect will be filtered out by the time we sample z(t) and it is passed through the ML detector.  Is this assumption correct?==
-. Calculate the expected value of x[k1] x[k2]. It is given by E{A^2} and is time independent (a special case of E{x[k1] x[k2]} = E{x[k1] x[k1 + m]} with m = 0.
+F3. No. Please refer to Fig. 3.13 in the text. Note that a1(t) is the integral of s1(t). Likewise a2(t) is the integral of s2(t).
-This x[k] is a WSS process.