1
$\begingroup$

Constant False Alarm Rate (CFAR) is an automatic thresholding technique. It's situated after the FFT processing in other to detect target peak. This probability of detection depends on clutter and noise level. In automotive RADAR, there are many other issues, like multi-path and interference. My questions are:

Is CFAR or XX-CFAR stay efficient detectors in presence of multipath and interference? If No, could you suggest me other techniques/methods to efficiently detect true targets in presence of interference and noise?

Improve this question
$\endgroup$
1
  • $\begingroup$ "best" can only be determined with respect to a given metric. Is your metric least missed targets? If so, which kind of targets, in which distances? I'm afraid this will really end up being your job, to model what you need to detect and apply it to the different detector methods and derive (or simulate) what works best. $\endgroup$
    – Marcus Müller
    Commented Aug 15, 2018 at 16:36

1 Answer 1

Reset to default
3
$\begingroup$

The "best" detector is a highly subjective subject, and there is (in my opinion) not a definitive answer here. I work in radar processing and I've used everything that you've mentioned in one form or another. All of these CFAR methods are tools, and all of them have good/bad applications. For instance, asking what the "best" wrench is in the toolbox highly depends on what you're going to use that wrench for. This is no different. OS CFAR can help with non-stationary noise floors which are varying quite a bit from CUT (cell under test) to CUT, but sometimes CA-CFAR can work just fine and is much simpler to implement.

In summary, they're all tools, and therefore no single CFAR detector is inherently "better" than any other one; it all depends on your system, the noise floor, the interference environment, the clutter environment, etc.

Improve this answer
$\endgroup$

Your Answer

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

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