0
$\begingroup$

I was successfully able to collect some CSI data using the existing tool(s) on GitHub (https://github.com/StevenMHernandez/ESP32-CSI-Tool). The CSI data is a pair of imaginary and real number which can use to extract amplitude and phase information. However, I have no prior knowledge of digital signal processing and I'm not sure how can to preprocess the CSI data to remove environmental noise like the Butterworth filter and/or PCA. I'm using Python and would like to know if this is an appropriate tool.

I'm trying to recreate steps from the following papers related to human activity detection using Wi-Fi CSI:

Improve this question
$\endgroup$
6
  • $\begingroup$ CSI doesn't include noise information; it describes the linear channel, not the additive noise. Also, not quite sure what you want to remove noise from – the CSI is not the signal. None of what you write makes too much sense. Maybe you need to take a step back and describe from a higher-up perspective what you want to do and achieve; I doubt what you're considering so far takes you any closer to that goal. $\endgroup$
    – Marcus Müller
    Commented Apr 11, 2021 at 12:14
  • $\begingroup$ Thank you for your time, I'm trying to recreate the steps mentioned in papers relating to activity detection using Wi-Fi CSI. Papers suggest after collecting CSI it's good to pass it through Butterworth filter to remove high-frequency noise. $\endgroup$
    – RikeshMM
    Commented Apr 11, 2021 at 12:42
  • $\begingroup$ which papers? which steps? I can't imagine a single case where Butterworth filtering on OFDM taps (i.e. CSI estimates) makes any sense for a time-variant channel, so I'd really recommend you link to the papers here. $\endgroup$
    – Marcus Müller
    Commented Apr 11, 2021 at 12:43
  • 1
    $\begingroup$ I have edited my question to support my question :) $\endgroup$
    – RikeshMM
    Commented Apr 11, 2021 at 12:51
  • 1
    $\begingroup$ "I have no prior knowledge of digital signal processing" I highly recommend acquiring "Digital Signal Processing in Modern Communication Systems" by Schwarzinger. It's inexpensive, very approachable, and will give you some of the backdround you need. $\endgroup$
    – MBaz
    Commented Apr 11, 2021 at 14:34

0

Reset to default