This question is specifically directed at Maximilian Matthé (as you work with 5G systems), and to anyone who can shed some light.

I am trying to implement a GFDM system with a total of N = K*M symbols, where K = total number of carriers, and M = total number of subsymbols. I believe that in order to improve the Out Of Band radiation only a subset of carriers, Kon < K, are allocated to carry information while the other carriers are left as blanks.

Therefore, when mapping the data symbols (that are QAM modulated) of the allocated carriers into a 2D time frequency matrix of the size KxM, there are zeros introduced into the modulated data. This results in an extra symbol at the origin of the QAM constellation.

This seems incorrect as there is another symbol that is part of the constellation. For example, 16-QAM would now have an additional symbol at the origin as shown in the image below.

enter image description here

However, I feel like this might be ok because the extra symbol is at the origin, and therefore has no amplitude, phase, and surely would not require any transmission energy.

Although, on the receiver side the constellation looks like this:

enter image description here

So now the added symbols at zero seems to have some amplitude and phase. Would they be taking up some energy ?

So my question is, is this ok? It is justifiable at all? Or should i rather skip out adding these zeros ?

  • 1
    $\begingroup$ Silence will be treated as zero anyways. The noise is coming from the receiver, which is expected. The only problem is that if position of zeros is not known at receiver a priori, the BER of this constellation will be worse than the original. $\endgroup$ – learner Sep 3 '17 at 16:27
  • 1
    $\begingroup$ Are you modulating the carrier that is on 0 Hz in baseband? Also, I'm not quite certain where in your signal processing these constellation diagrams occur? As far as I see, you'd do a constellation diagram for each of your used subcarriers, but not for those you don't use? $\endgroup$ – Marcus Müller Sep 3 '17 at 19:15
  • 1
    $\begingroup$ If I understand well, "M = total number of subsymbols" is QAM alphabet size ? "in order to improve the Out Of Band radiation ... Kon < K" means you use some null/virtual carriers, i.e. FFT size is greater than the number of used subcarriers ? If it is the case, this is exactly what have been done in commercial 4G LTE. New phase/amplitude are due to noise and/or channel equalization step. What you mean "new energy" are energy of noise at receiver and it is expected. Like @learner have said, you should have known the positions of null subcarriers then just ignore them. $\endgroup$ – AlexTP Sep 3 '17 at 19:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.