From kozick@eg.bucknell.edu Mon Feb 15 13:50:32 1999
Date: Mon, 15 Feb 1999 13:50:19 -0500 (EST)
From: Kozick Rich  <kozick@eg.bucknell.edu>
To: armentrt@bucknell.edu, baxter@bucknell.edu, bixby@bucknell.edu,
        rbrooks@bucknell.edu, jbullard@bucknell.edu, campbell@bucknell.edu,
        daub@bucknell.edu, everitt@bucknell.edu, gamache@bucknell.edu,
        afgibson@bucknell.edu, bjgreen@bucknell.edu, sgunnels@bucknell.edu,
        hajdu@bucknell.edu, louie@bucknell.edu, onybrchi@bucknell.edu,
        phan@bucknell.edu, rudis@bucknell.edu, senator@bucknell.edu,
        cwil@bucknell.edu, witherup@bucknell.edu, yyoon@bucknell.edu,
        blixt@bucknell.edu, ganiear@bucknell.edu, goossen@bucknell.edu,
        sgordon@bucknell.edu, keeney@bucknell.edu, dkline@bucknell.edu,
        mayega@bucknell.edu, mmccrthy@bucknell.edu, kmehaffy@bucknell.edu,
        michael@bucknell.edu, eneff@bucknell.edu, pankake@bucknell.edu,
        pickert@bucknell.edu, pribik@bucknell.edu, reed@bucknell.edu,
        dseiler@bucknell.edu, sweet@bucknell.edu, vallone@bucknell.edu,
        vary@bucknell.edu, ryoung@bucknell.edu, rejmer@bucknell.edu,
        welliver@bucknell.edu, zhong@bucknell.edu, kozick@bucknell.edu
Subject: ELEC 471 HW 4 notes

Hi,

Below are some notes regarding homework 4 for ELEC 471.  These were
mentioned in class, but I want to make them clear to everyone.

1.  No problems from Chapter 2 are due on Feburary 18.

2.  A sample MATLAB program is available at 
    http://www.eg.bucknell.edu/~kozick/elec471/digcom1.m

    This is the program that we developed in class on February 11.

3.  You should ANALYZE the probability of a bit error for the system
    that sends each bit three times and takes the majority decision as
    the overall decision.  You should do the analysis for the general
    case of a RANDOM decision when x = 0 is received.  Your analytical
    result should agree closely with your simulation result.

4.  Be sure to do enough runs for each simulation.  If the expected
    probability of a bit error is 0.01, then you should do at least
    1000 (or 10,000 is better) runs in order to "see" enough errors.

5.  Here is a MATLAB hint for taking the majority of three decisions:
    Do a loop over the three transmissions, and make a decision on
    each of the three bits:

    for p=1:3  % Loop over three transmissions

    % Decision logic ...

    shat(p) = 0 or 1 ...   % Stores the three bit decisions in an array

    end

    Then the following will take the majority of the three individual
    bit decisions:

    majority = median(shat);

Feel free to see me if you have any questions this week.

Thanks.

-- Rich Kozick