Robust Energy Efficient Cooperative Spectrum Sensing in Cognitive Radios
George Atia, Erhan Baki Ermiş, Venkatesh Saligrama · 2007 IEEE/SP 14th Workshop on Statistical Signal Processing · 2007
A crucial task for a network of cognitive radios is to detect occupied frequency bands, to protect transmissions of primary users, and to identify spectrum holes (unoccupied bands) to maximize the utilization of wasted resources. This paper is motivated by the need to account for challenging constraints that naturally arise in such applications such as channel model uncertainties and demanding sensitivity constraints of the sensing devices. We propose a False Discovery Rate (FDR) based cooperative strategy to sense the occupancy of the spectrum. The proposed strategy is robust to significant uncertainties such as lack of CSI, fading and shadowing effects. Furthermore, it is shown that the cooperative sensing strategy significantly reduces sensitivity requirements. We quantify the effect of channel occupancy rate on the required cooperation degree for achieving a guaranteed level of primary user protection.