0
$\begingroup$

I'm trying to implement a tracking system (undergrad thesis) using DOA of acoustic signals. I'm testing this by playing a 2kHz sound and receiving the signal with four microphones in a ULA.

So, the band I'm interested in is narrowband but naturally the signal received is wideband.

I have been able to successfully implement a time delay beamformer however when I have tried to implement a phase shift beamformer I have failed.

My suspicion is that you can only perform phase shift beamforming techniques on narrowband signals and will not work even if you are looking for a narrowband within a wideband signal. Am I correct?

If so, will a simple filter be good enough to correct this issue to transform the signal into a narrowband one?

|improve this question
$\endgroup$
0
$\begingroup$

My suspicion is that you can only perform phase shift beamforming techniques on narrowband signals and will not work even if you are looking for a narrowband within a wideband signal. Am I correct?

No; the correlation coefficient of your reference signal to your received signal should, given additive, phase-uniform, white noise, have an expectation value of the phase shift the individual receiver sees, no matter how wide your band is.

Note that this is only true if your transmitter/receiver system can be modeled as linear phase for the bandwidth we're talking about, but without loss of generality, radar/sonar/sound systems are designed to pretty much have exactly that characterestic.

|improve this answer
$\endgroup$
  • $\begingroup$ i'm not using a reference signal, just getting the power of the output... $\endgroup$ – Makoto Aug 21 '16 at 19:02
  • $\begingroup$ well, that's not phase estimation then, is it $\endgroup$ – Marcus Müller Aug 21 '16 at 19:05
  • $\begingroup$ maybe not, fairly new to this, i initially implemented a delay and sum beamformer, which worked using a time delay implementation but stopped working when i tried implementing it with a standard steering vector. i then tried using matlab's phased array object, and it worked for the time delay beamformer but couldnt get it to work with the phase delay beamformer. but i havent seen any mention of a reference signal. $\endgroup$ – Makoto Aug 21 '16 at 19:14
  • $\begingroup$ sorry, I should have clarified that: I meant to say "pick any of your receiver elements, declare it as your reference, correlate against other elements' signals, and you should get complex correlation coefficients that include angle of arrival info" $\endgroup$ – Marcus Müller Aug 21 '16 at 19:16
  • $\begingroup$ I'm a little confused about you implementing it with a steering vector: That vector is what you want to estimate if you do a DOA estimator, isn't it? $\endgroup$ – Marcus Müller Aug 21 '16 at 19:17

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.