ELEC 470, Spring 2004
Prof. Rich Kozick

ELEC 470: Project 2

Recovery of Message Signals from a Frequency Division Multiplexed Mystery Signal


Date Assigned: Tuesday, April 6, 2004
Date Due: Tuesday, April 20 and Thursday, April 22, 2004


The Problem:

Because of your expertise in analog communication systems, you have been hired by a special agent to recover some messages from a sampled signal. Frequency division multiplexing (FDM) with some form of analog modulation (AM, DSB-SC, SSB, VSB, FM) was used to encode the messages to produce the signal that is available to you. Your job is to analyze this signal, determine how many messages are present, determine the carrier frequencies, determine the type of modulation used at each carrier frequency, and finally recover each of the individual messages through an appropriate demodulation process.

Some Things That Are Known:

  1. More than one message signal is present.
  2. The message signals have zero DC (or average) level.
  3. The bandwidth of each message signal is less than 200 Hz.
  4. Each carrier frequency is a multiple of 1000 Hz.
  5. A different type of modulation may be used at each carrier frequency.

Some Hints:

  1. The sampled signal that you are to analyze can be loaded into MATLAB by first saving the file data in your account with the filename data, and then typing at the MATLAB prompt:
    load data -ascii
    x = data;
    clear data
    The variable x contains the signal to be processed.
  2. The spacing between samples in this file is 50 microseconds, which corresponds to a sampling rate of 20,000 samples per second.
  3. The FFT is a useful tool for determining the carrier frequencies and the modulation type.
  4. The example MATLAB program mod_ex.m shows how to use the FFT, perform various types of demodulation, and perform low-pass and high-pass filtering. Note that the sampling rate in mod_ex.m is different from the sampling rate in this project.
  5. If you need a coherent detector, then the phase of the local oscillator will be important. Think of some (simple?) ways to determine the proper phase. (Hint: The phase shift must lie in the range from 0 to 2*pi radians.)

What to Submit and When:

  1. For Tuesday, April 20, please come to class with questions that you have, and I will help you. You should have a very good idea by then about which types of modulation are present, but you may have questions about how to do certain processing in MATLAB. In other words, by April 20 you should have analyzed the available signal and developed a system-level description of the processing that is needed to recover the messages. I will help you to implement your system in MATLAB.
  2. Submit your results on Thursday, April 22. Your results should include a discussion of your approach, including how you determined the modulation types, how you performed the processing, the MATLAB program that you used, and plots of the demodulated message signals in the time and frequency domains.
  3. Your report should be mostly explanation, and not a MATLAB program with some plots attached!
Thank you and have fun!