Wideband spectrum sensing for cognitive radio via phase-field segmentation

In this paper, we investigate the problem of wideband spectrum sensing for detecting vacant frequency subbands in opportunistic cognitive radio. To this end, we identify the irregularities (discontinuities) in the estimated power spectrum density. More precisely, we propose a new mathematical framework based on phase-field segmentation method, classically used in the image processing community. We show that by properly setting the parameters of the phase-field functional, robustness to fluctuations of the threshold value (caused for instance by estimation errors) used for spectrum sensing can be achieved. Our numerical results indicate that the sensing accuracy is improved, while the computational complexity is reduced, when compared to conventional spectrum sensing methods.