<< Chapter < Page Chapter >> Page >

The default settings in dae.m are used to perform the equalizer designs for three channels.The source alphabet is a binary ± 1 signal. Each channel has a FIR impulse response, and its outputis summed with a sinusoidal interference and some uniform white noise before reaching the receiver.The user is prompted for

  1. choice of channels (0, 1, or 2),
  2. maximum delay of the equalizer,
  3. number of samples of training data,
  4. gain of the sinusoidal interferer,
  5. frequency of the sinusoidal interferer (in radians), and
  6. magnitude of the white noise.

The program returns plots of the

  1. received signal,
  2. optimal equalizer output,
  3. impulse response of the optimal equalizer and the channel,
  4. recovery error at the output of the decision device,
  5. zeros of the channel and the combined channel–equalizer pair, and
  6. magnitude and phase frequency responses of the channel, equalizer, and the combined channel–equalizer pair.

For the default channels and values, these plots are shown in [link][link] . The program also prints the condition number of R ¯ T R ¯ , the minimum average squared recovery error (i.e., the minimum value achieved by the performance functionby the optimal equalizer for the optimum delay δ o p t ), the optimal value of the delay δ o p t , and the percentage of decision device output errors inmatching the delayed source. These values were as follows:

  • Channel 0
    • condition number: 130.2631
    • minimum value of performance function: 0.0534
    • optimum delay: 16
    • percentage of errors: 0
    Trained least-squares equalizer for Channel 0: Time responses. The received signal is messy and cannot be used directly to recover the message. After passing through the optimal equalizer, there is sufficient separation to open the eye. The bottom left figure shows the impulse response of the channel convolved with the impulse response of the optimal equalizer, it is close to an ideal response (which would be one at one delay and zero everywhere else). The bottom right plot shows that the message signal is recovered without error.
    Trained least-squares equalizer for Channel 0: Time responses. The received signal is messy and cannot be used directly torecover the message. After passing through the optimal equalizer, there is sufficient separation to open the eye.The bottom left figure shows the impulse response of the channel convolved with the impulse response of the optimalequalizer, it is close to an ideal response (which would be one at one delay and zero everywhere else). The bottom right plotshows that the message signal is recovered without error.
    Trained least-squares equalizer for Channel 0: Singularities and frequency responses. The large circles show the locations of the zeros of the channel in the upper left plot and the locations of the zeros of the combined channel–equalizer pair in the lower left. The *** represents the frequency response of the channel, — is the frequency response of the equalizer, and the solid line is the frequency response of the combined channel–equalizer pair.
    Trained least-squares equalizer for Channel 0: Singularities and frequency responses. The large circles show the locations of thezeros of the channel in the upper left plot and the locations of the zeros of the combined channel–equalizer pair in the lower left.The *** represents the frequency response of the channel, — is the frequency response of the equalizer, and the solid lineis the frequency response of the combined channel–equalizer pair.
    Trained least-squares equalizer for Channel 1: Time responses. As in Figure 13-9, the equalizer is able to effectively undo the effects of the channel.
    Trained least-squares equalizer for Channel 1: Time responses. As in [link] , the equalizer is able to effectively undo the effects of the channel.
    Trained least-squares equalizer for Channel 1: Singularities and frequency responses. The large circles show the locations of the zeros of the channel in the upper left plot and the locations of the zeros of the combined channel–equalizer pair in the lower left. The *** represents the frequency response of the channel, — is the frequency response of the equalizer, and the solid line is the frequency response of the combined channel–equalizer pair.
    Trained least-squares equalizer for Channel 1: Singularities and frequency responses. The large circles show the locations of thezeros of the channel in the upper left plot and the locations of the zeros of the combined channel–equalizer pair in the lower left.The *** represents the frequency response of the channel, — is the frequency response of the equalizer, and the solid lineis the frequency response of the combined channel–equalizer pair.
    Trained least-squares equalizer for Channel 2: Time responses. Even for this farily severe channel, the equalizer is able to effectively undo the effects of the channel as in Figures 13-9/> and 13-11/>.
    Trained least-squares equalizer for Channel 2: Time responses. Even for this farily severe channel, the equalizer is able to effectivelyundo the effects of the channel as in Figures [link] and [link] .
    Trained least-squares equalizer for Channel 2: Singularities and frequency responses. The large circles show the locations of the zeros of the channel in the upper left plot and the locations of the zeros of the combined channel–equalizer pair in the lower left. The *** represents the frequency response of the channel, — is the frequency response of the equalizer, and the solid line is the frequency response of the combined channel–equalizer pair.
    Trained least-squares equalizer for Channel 2: Singularities and frequency responses. The large circles show the locations of thezeros of the channel in the upper left plot and the locations of the zeros of the combined channel–equalizer pair in the lower left.The *** represents the frequency response of the channel, — is the frequency response of the equalizer, and the solid lineis the frequency response of the combined channel–equalizer pair.
  • Channel 1
    • condition number: 14.795
    • minimum value of performance function: 0.0307
    • optimum delay: 12
    • percentage of errors: 0
  • Channel 2
    • condition number: 164.1081
    • minimum value of performance function: 0.0300
    • optimum delay: 10
    • percentage of errors: 0

Get Jobilize Job Search Mobile App in your pocket Now!

Get it on Google Play Download on the App Store Now




Source:  OpenStax, Software receiver design. OpenStax CNX. Aug 13, 2013 Download for free at http://cnx.org/content/col11510/1.3
Google Play and the Google Play logo are trademarks of Google Inc.

Notification Switch

Would you like to follow the 'Software receiver design' conversation and receive update notifications?

Ask