Robust Multiband Spectrum Sensing in Cognitive Radio Networks via Exponentially Embedded Family Criterion

This paper proposes a robust spectrum sensing scheme for cognitive radio networks. The proposal performs spectrum sensing simultaneously over the total channels rather than a single channel at a time. It first estimates the number of occupied channels via the exponentially embedded family criterion, and then determines the occupancy status for each channel by sample power. The proposed method is robust to noise uncertainty since noise power estimation is avoided. Simulations demonstrate that our approach is robust to noise uncertainty and outperforms existing sensing methods. 