After a chirp is detected, it is filtered out from other sounds
and overlapping chirps that might also be present in the selected
time segment. For this purpose, initially the time segment from
all of the microphone signals are passed through a band-pass
filter to remove the unwanted low and high frequency noises
that are outside the chirp’s frequency range (see Figure 8.b).
Furthermore, Fractional Fourier transform (FRFT) [37] is
used to remove noise within the frequency range of the chirp.
Unlike Fast Fourier transform (FFT), the FRFT provides a
compact representation of chirp signals, which makes it possible
to remove the noise inside the same frequency region as
the chirp, that cannot be removed with traditional frequency
domain filters.