In the area of broadband spectrum sensing, there are multiple methods used to detect energy. Two common methods are:

  1. Energy detection in the time domain (matlab example)
  2. Spectrogram (often defined by a time and frequency resolution [Fs/NFFT], overlap percentage, and number of noncoherent averages [NAVG])

When doing energy detection in the time domain, a polyphase channelizer is often used make smaller subchannels of a large chunk of spectrum.


  1. How do you compare the sensitivity (ie probability of detection) between the two methods?
  2. Is it accurate to say that the difference between the detection SNRs would be: $$ \Delta \text{SNR} = 10\log_{10}(\Delta f_1 / \Delta f_2) + 10\log_{10}\sqrt {N_{AVG}}$$ where $$ \Delta f = 1/F_s $$ for option 1 and $$ \Delta f = F_s/N_\text{FFT} $$ for option 2?
  • $\begingroup$ 1. by modelling or simulating signals of different powers and deriving or measuring least reliable detection (you'll need to define a least probability of detection). $\endgroup$ – Marcus Müller Feb 26 '20 at 19:34
  • $\begingroup$ 2. no, since that depends on your signal model. $\endgroup$ – Marcus Müller Feb 26 '20 at 19:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.