Adaptive Channel Equalization Help Needed!!!!

[lassmat]

Hi!!
I'm doing a project on Channel Equalization.I have the input signal to the channel which is also the reference signal.I also have the output signal from the channel which is the input to the adaptive filter.

Now I have to delay the reference signal by various values and plot those values with error signal from the output of the adaptive filter i.e I'm supposed to plot error on the y-axis versus delay on the x-axis.

I'm also supposed to plot error on the y-axis and filter length (adaptive filter length) on the axis.

The reason to plot these plots is to determine the optimal delay and optimal filter length.

How do I go about this? I will do it in matlab.

Your help will be appreciated.

Regards

 

[priyankguddu]

Hi

Just some clarifications.. first of all isn't it too ideal to have the input signal itself as the reference signal (cos if you know the signal, whats the use of a filter there). Second thing is: you are saying that you would plot the error signal between the input and delayed version of the refernce (which is the input signal). are you considering a non-dispersive channel.

 

[lassmat]

I think I have to make things clearer.I have been given two signals, say, x and
r. x is the original signal and r is the output of the CHANNEL which is a corrupted version of x.

r is to be the input of the equalizer which will be driven by various algorithms (LMS, Leaky LMS, RLS etc ).

Lets call the out of the qualizer (driven first by LMS) be y.I suppose the difference between y and a delayed version of x is the the error, e.

Now, the channel is completely unknown.Also, I have no idea about the equalizer tap weights.

The idea of the project is to use those two signals, x and r, and the algorithm to come out with plots of (1) e vrs delay (2) e vrs filter length.

The reason of these plots is to determine the Optimal Delay and Optimal filter length for the algorithm.

I hope it a bit clear now.I really need help because I'm completely new to Adaptive filters.

I hope to hearing from you.

Regards

 

[just4me]

This is a very simple project and it becomes simpler if using MATLAB Simulink. I had done the same thing. There is a simulink simulation called "lmseq". Try running it. It does the same thing that you are asking for. You will understand the concept better by running the simulation a couple of times. It adjust the filter weights to minimize the output given by the LMS filter.

 

[lassmat]

I really appreaciate your reply.I'll try that method and see.But this is not a simulation project.I have to write the matlab code myself.I'll definitely come back to you since you done that before.

Regards