I am looking for a method to make a multi-user channel estimation with CDMA.

Here is my question:

Is there a way to estimate multi-user channels with their spreading codes? If yes, what is the method?


1 Answer 1


Sure. The idea behind CDMA is that different users codes are orthogonal.

So, since your codes probably are pretty white, meaning that their autocorrelation function is essentially a dirac delta, you can simply encode an pulse for every user (or by every user), and use the standard correlation receiver to despread and thus recover each individual channel impulse response.

  • $\begingroup$ should it be CDMA instead of OFDM? $\endgroup$
    – AlexTP
    Dec 17, 2018 at 14:00
  • $\begingroup$ ah i shouldn't be answering different questions on different sites simultaneously, @AlexTP $\endgroup$ Dec 17, 2018 at 14:11
  • 2
    $\begingroup$ (don't tell anybody, but OFDMA is just a special case of CDMA with very specific spreading sequences ;) ) $\endgroup$ Dec 17, 2018 at 14:11
  • 1
    $\begingroup$ Yes! A 64 point DFT gives you 64 spreading codes with complex coefficients. Just like the 64 Walsh matrix gives you 64 codes using only +1/-1 coefficients. The 2pt DFT and the 2x2 Walsh matrix are identical. $\endgroup$ Jan 11, 2020 at 17:39
  • 1
    $\begingroup$ … and in the end, your despreading is naught but a matched filter sampled at appropriate times. $\endgroup$ Jan 11, 2020 at 18:55

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.