If we have a non-linear function $f$, that is applied to input $x$, we have then the output $y=f(x)$

Using Bussgang decomposition we can linearize this nonlinearity and express $y$ as $y=Bx+ η$,

Where B could be found as per the below formula below:

$B= \frac{E(yx^{*})}{E(|x|^2)}$

We can use input pilots at the beginning, so assume we we know $x,y$ and we get $B$, from the just above formula.

Then we obtain $\eta= y-Bx$, and this was for the training phase.

But now for the testing , I already have $B$ that I obtained previously from the training phase, but I don’t have $\eta$ nor $x$, so how can I find back the input $x$?

  • $\begingroup$ Something sounds strange. $B$ is obtained as the ratio of two expectations (probably, found empirically on many trials, not just on a single one) to minimize the expected square error, so what "input $x$" are you trying to find? $\endgroup$
    – fedja
    Aug 31, 2020 at 1:09
  • $\begingroup$ @fedja yeah B is the ratio of two expectations, i did mention the formula in the question. Concerning the input x it is the variable of linearization, may be i am not understanding Bussgang properly, i would appreciate if you can refer me a good source code imlementation, or resource to understanding Bussgang decompostion better :) thanks in advance $\endgroup$
    – e. sfe
    Aug 31, 2020 at 17:43


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Browse other questions tagged or ask your own question.