On the use of blind source separation for peak detection in spectrum sensing

Applying wavelet edge detection technique on observed wideband spectrum, results in a signal which contains frequency band boundaries information. Resultant signal contains peaks at locations corresponding to frequency band boundaries i.e. start and end locations of frequency bands. In the presence of noise resultant signal contains mixture of true peaks and noisy peaks. A threshold value is required to extract true peaks efficiently from mixture. In this paper calculation of threshold value is performed using blind source separation technique. Probability of detection and success ratio plots are used to evaluate proposed technique. Success ratio plot shows improvement of 4 dB and probability of detection plot shows improvement of 8 dB. Moreover, the proposed algorithm is based on the received signal and does not require any apriori information.