0

I am working with ultrasonic transducers with frequencies 100 kHz, one for transmission one for reception. The transmitted signals are chirps. I would like to measure the time of flight between transmission and reception. Can anyone help me with how could I do the phase/amplitude equalization of impulse response of transducer? My received signal is distorted and attenuated.

Deanna77
  • 41
  • 3
  • This is very helpful for understanding the equalization. However, I want to measure time of flight between transmission and reception. As with equalization I would remove the group delay, I would also remove the time of flight that I want to measure. Am I right? – Deanna77 Jun 24 '20 at 21:57
  • 1
    (You should update your question since it does not indicate that important point) I have another post using the same process but specific on measuring the delay- let me link that for you as well. – Dan Boschen Jun 24 '20 at 21:58
  • 1
    See this: https://dsp.stackexchange.com/questions/63141/how-determine-the-delay-in-my-signal-practically/63221#63221 and https://dsp.stackexchange.com/questions/63759/a-delay-between-two-filtered-chaotic-signals/63761#63761 – Dan Boschen Jun 24 '20 at 22:00
  • @DanBoschen I have looked into the links you have suggested me, but I still struggle understanding the following: If I want to measure time of arrival(distance) between transmission and reception, for different mediums in between transmitter and receiver, if I do the equalization, then I would have estimation of impulse response not only of my transmitter and receiver, but from the medium part as well. And wouldn't then the equalization process nulify the delay that I want to measure? I would like somehow remove the distortions that happen due to characteristics of transmitter and receiver. – Deanna77 Jun 25 '20 at 20:03
  • Instead of equalizing the channel, you would be using the equations and process to estimate the channel (in a least squares sense). The posts detail how to do this by simply swapping Tx and Rx. From the channel estimate you can get the frequency response of the channel and from that you can accurately estimate the delay which may be frequency dependent (this process will confirm that). If you find that the channel is non-dispersive (constant group delay), then you can safely used a matched filter/ cross correlation technique—- and the process I outline confirms that – Dan Boschen Jun 25 '20 at 20:27
  • So the link I pointed you in the comment directly above to was very specific to this and refers to measuring the delay and the channel and not equalizing it; I want to be sure you were looking at that specifically – Dan Boschen Jun 25 '20 at 20:29
  • But as I understood from the post, with the LMS technique I would estimate the channel and then filter my signal with the estimated equalization filter coefficients. However, from you comment now I understood that I would be estimating the channel impulse response in order to see if the phase is linear, and if it is, than I can use matched filtering. And what if it is not? I am sorry for repeating multiple times, but I would like to be sure that you understand what I am trying to accomplish. – Deanna77 Jun 25 '20 at 20:50
  • This post does not show that process: https://dsp.stackexchange.com/questions/63141/how-determine-the-delay-in-my-signal-practically/63221#63221 In equalization you are not filtering with an estimate of the channel but you are filtering with coefficients that would optimally undo what the channel has done. In the link above I show the approach that simply estimates the channel, no further filtering. – Dan Boschen Jun 25 '20 at 20:53
  • I believe that linked post completely answers your question - we should close this one out as they discourage extended discussion and ask further questions there if any of it isn't clear but it is specifically how to get the time of propogation. (Also the question here is still asking about equalization which as you now understand isn't even what you are looking to do) – Dan Boschen Jun 25 '20 at 20:56
  • Let us continue this discussion in chat. – Deanna77 Jun 25 '20 at 20:58

0 Answers0