0
$\begingroup$

I need to calculate BER (bit error rate) for a system using spreading spectrum DS over ultra wideband (UWB). Modulation is bpsk. The spreading sequence lengths are not equal for all users. Receiver is matched filter. I know the equations of DS-UWB transmission signal and template signal which I should use at receiver to extract the user's data. I know that I should use convolution between the received signal (transmitted + AWGN noise) and template signal. My problem is, how to calculate the result of that convolution? is it integration or multiplication? thanks in advance

$\endgroup$
  • $\begingroup$ It's not so clear what you're asking. Convolution is...convolution. It involves multiplicaton and integration. $\endgroup$ – Deve Jun 29 '13 at 10:45
  • $\begingroup$ @Deve thanks for response. Can you tell me some reference where I can find details how mathematically the convolution output is calculated for ds-uwb? $\endgroup$ – Mohammed Jul 1 '13 at 9:49
  • 1
    $\begingroup$ I'm afraid I can't. What makes the convolution in DS-UWB different from a conventional convolution? $\endgroup$ – Deve Jul 1 '13 at 11:50
  • $\begingroup$ @Deve Actually this is my problem, I don't know how to calculate convolution between two sequences with different lengths. $\endgroup$ – Mohammed Jul 1 '13 at 12:46
1
$\begingroup$

Assume two discrete-time sequences: $x$ (length $L_\mathrm{x}$) and $y$ (length $L_\mathrm{y}$). Furthermore, $L_\mathrm{x} \geq L_\mathrm{y}$. To calculate the discrete-time (linear) convolution of $x$ and $y$, first pad $y$ with $L_\mathrm{x} - L_\mathrm{y}$ zeros, i.e. $$ y'(n) = \begin{cases} y & 0 \leq n < L_\mathrm{y}\\ 0 & L_\mathrm{y} < n \leq L_\mathrm{x} - 1 \end{cases} $$ Then calculate the convolution of $y'$ and $x$ as described here, for example.

$\endgroup$
  • $\begingroup$ The zero-padding is not required. It will just increase the amount of multiplications and since most of them will have one factor of zero, nothing is gained. $\endgroup$ – jan Aug 1 '13 at 9:55

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.