Applying constant false alarm rate is a critical step for estimation of performance of all detectors. Has anybody experience on practical implementation of this estimator for energy detector? Can anybody explain and introduce a code to me?
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
2
Constant false alarm rate detection is more of a general technique that underlies a class of approaches; you can't really say that there's "an" implementation of CFAR processing. CFAR techniques are useful when you need to detect the presence of a signal where there is some uncertainty as to what the magnitude of the signal will be, especially relative to any background noise or interference. The general principle of constant false alarm rate detection is that the detector makes some effort to estimate the background level and set a detection threshold adaptively relative to the observed ambient environment. This approach makes the performance of the detector (typically measured quantitatively using the probability of detection $P_d$ and probability of false alarm $P_{fa}$) insensitive to the bulk magnitude of the signal under observation; only relative changes in the statistics of the signal over time (such as a transition from noise only to signal-of-interest plus noise) are important.
-
$\begingroup$ Thanks and I should add that it has several types. CA-CFAR is one of the types. Now could you please let me know how in practice people simulate this? Indeed in practice means that without using statistical distribution how based on concept of false alarm we can simulate CFAR detection. $\endgroup$– HosseinCommented Dec 13, 2011 at 22:25
-
$\begingroup$ I'm not sure what you're asking. If you're familiar with how radar systems work, then implementing cell-averaging CFAR detection is very straightforward. If you're going to simulate it, then you need to have some kind of statistical model for the environment and radar measurements. Otherwise, how do you generate the data for the simulation? $\endgroup$– Jason RCommented Dec 15, 2011 at 14:45