0
$\begingroup$

In order to detect the presence of a signal impaired by noise, what is the best technique (best metrics) one can use ?

  1. energy detection,
  2. autocorrelation,
  3. matched filter based sensing,
  4. CRC data aided or not .. an explanation of the underlying algorithms would also be appreciated
$\endgroup$
  • 1
    $\begingroup$ 1. If you know the signal 'a priori' then 'correlation' is the mathematical principle to be used, 'matched filtering' is the implementation of the correlation process. 2. These are all vast subjects, depending on the field in which you want to implement this, we will able to narrow it down for you. $\endgroup$ – abhilash May 4 at 6:50
1
$\begingroup$

There is no best rule, it depends on what we know about the signal.

Energy Detector: The energy Detector makes a decision on presence or absence of a signal based on sum of squared samples. This rule comes from the fact that the signal to be detected is inherently assumed to be random following a wide sense stationary process with a known PDF, the likelihood ratio test then simply determines this rule.

Matched filtering and autocorrelation: The matched filtering is nothing but the correlation of the. Received samples with the known signal itslef, so they are inherently the same, the difference compared to energy Detector is that here the signal is assumed to be a known deterministic quantity. The matched filter is the optimal detector in low SINR scenarios.

CRC data: A check on CRC is an agreement between the receiver and tarsmitter in whether an error has occurred or not while transmitting data on a wired/optical/Wireless channels. By itself it is not a signal detection scheme. It only detects whether an error has occurred or not. Once an error is detected the receiver can then request the trasnmitter to send the data again by sending a NACK (not acknowledge) signal to the trasnmitter (there are other ways to handle it as well, like soft combining etc.)

| improve this answer | |
$\endgroup$

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.