Hi there!
I'm a student in telecommunication and I'm currently working on a virtual ditigal communication lab for a training course. I must produce rectangular and raised-cosine shaped signal, add noise and filter them with a matched filter, display the eye diagram, and retrieve the symbols I generated.
This is not real time signal processing (the machine is too slow), so the shape of my symbols are stocked in vectors (one for real part, the other for the imaginary). I was wondering howto implement the matched filter?
A very basic assumption was to create a for boucle that will dissassemble the vector, whith 3 entries : one for the impulse shape, the others for real and imaginary noised signals. in the boucle, I would use ppp-fourier tranforms, multiply the noised signals by the reference and use reverse fourier transformation. Next, a matlab script would extract from the reverse FT coefficients of the output signal (enclosed is the vi of my virtual lab, many function are still not good, but I'm waiting to finish them).
Does this behave as a matched filter?
If yes, I wonder how to analyse the output of the matched filter, since its output doesn't ressemble to the emitted symbols, and there is an overlapping between symbols... Does anyone could help on this? (there are no documentation on the net, and my tutor isn't very available to help me...)
Thanks
PS : If you were to send VIs, I'm using Labview 7.0 student version...
Message Edited by fenice on 06-23-2006 02:58 AM
