Recent Forum Posts
From categories:
page »

Hi all:

Good Luck for the final!

Gig'em aggies!

Good Luck to you all by ECEN661 StudentECEN661 Student, 10 May 2010 06:11

Dr. Narayanan

Could you please upload the notes and HW solutions before the test?

Thank you

We are also waiting for the homework solutions. Please upload them, is possible.

Dear Prof.Krishna Narayanan

When will you post the notes from mid-term? Thank you very much.

It is unlikely. If you want to talk to me about your performance, you can come by on Friday. I think I will do most of the grading only over the weekend or right after the finals.

Re: Project by Krishna NarayananKrishna Narayanan, 07 May 2010 02:15

Will we get to know our project scores before exam?

Re: Project by ECEN661 StudentECEN661 Student, 05 May 2010 22:01

How should we upload the m-files to the web? Moreover, should it be up to Tuesday that we turn in our reports?

Re: Project by ECEN661 StudentECEN661 Student, 03 May 2010 03:47

No, there should be no difference in the implementation. However, make sure you compute the branch metric properly as the square of the magnitude of the difference between what you observe at that time instant and what the noiseless output along that branch should be.

Re: Project by Krishna NarayananKrishna Narayanan, 03 May 2010 00:34

Is there any difference in the implementation of viterbi if the inputs and tap coefficients are complex and not real?

Re: Project by ECEN661 StudentECEN661 Student, 02 May 2010 01:53

The exact expression for the BER with a Viterbi algorithm is quite difficult to compute (there is no closed form expression or even an infinite series that can be numerically computed). However, we can obtain good lower bounds at high SNR. We can also obtain upper bounds. The relative easy bound to compute would be the lower bound based on computing the minimum Euclidean distance error event just like the example I worked out in class for three tap channel. There are some what easy ways to find this using the MATLAB program that you have already written for the Viterbi Algorithm, it is a bit difficult to explain this here.

As for the noise, there is often some confusion among the origin of what we call as noise in our model. This is thermal noise which has nothing to do with the individual paths that the transmitted signals take to get to the receiver. Read the section on thermal noise from Thermal noise and see that this noise occurs in the electronic circuitry at the receiver. So, you will NOT add noise to each path, but only once to the received signals.

Re: Project by Krishna NarayananKrishna Narayanan, 01 May 2010 13:12

What is the difference between the steps 5 and 6 in the task list? Are there two ways of computing BER - with using and without using Montecarlo simulations?

Re: Project by ECEN661 StudentECEN661 Student, 27 Apr 2010 19:21

How to check the performance of viterbi algorithm i.e. What is the theoretical expression for Pe vs SNR graph for viterbi algorithm?
Moreover, How is SNR (or Es/No) defined? Do we add noise to each channel (path) seperately or to the sum?

Project by ECEN661 StudentECEN661 Student, 23 Apr 2010 16:58

We need some office hours, discussing HW 7 before Tuesday . Is it possible for you?

HW 7 by ECEN661 StudentECEN661 Student, 23 Apr 2010 16:16

1. In Question no. 4, whats the definition of ISI? More precisely, do we quantify it in terms of power or something ?

Homework 6 by ECEN661 StudentECEN661 Student, 03 Apr 2010 19:46

I think that's why figure 4.6-1 is important.
The figure shows us the relaton btw data rate and SNR, given probability error.
So, when we consider SNR-Pe graph for two different schemes,
first find how much energy do we need for acheive the same Pe,
and then compare the data rate given those amount of SNR.
In this manner, one may be better in SNR but worse in Data rate than the other.
The decision what scheme you're going to use is up to you.

After we learned a lot of
SNR-Pe graph(Eb/No to error probability graph)
for different signaling scheme
the book summarize that
"One can compare them[different signaling scheme]
on the basis of SNR required to achieve a specified
probability of error. However, such a comparison would
not be very meaningful, unless it were made on
the basis of some constraint, such as a fixed data rate
of transmission or, equivalently, on the basis of a fixed

since we didn't talked about such constraints when we introduced
the SNR-Pe graph, are they making such assumption?
if not, how can we make use of it?

how can we equate data rate and bandwidth?
it make sense, but I don't feel very solid about it.

SNR-Pe graph useless? by ECEN661 StudentECEN661 Student, 26 Mar 2010 09:15

Hi David,

I tried Insert Image Wizard and for picture "URL". Here is an example…


I hope it works.

how do you guys insert image to this wiki?

I tried "insert image" and "insert image wizard" but failed
insert image need to filled in to the code

[[image source]]

but I don't know what should I filled into
and the wizard need URL of the image
I tried my image on picasa
but it takes a long time to upload which I don't know
if it will succeed or not

any ideas?


howto upload image? by ECEN661 StudentECEN661 Student, 15 Mar 2010 03:46

In problem 3 you should be able to write the answer in terms of product of Q functions. In Problem 5, you should be able to write the answer as a Q function


In Q3. and Q5., do we leave the probability of error in terms of a double integral?

page »
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License