A Near-Optimal LLR Based Cooperative Spectrum Sensing Scheme for CRAHNs

In Cognitive Radio Ad Hoc Networks (CRAHNs), cooperative spectrum sensing schemes exploit spatial diversity of the Secondary Users (SUs), to reliably detect an unoccupied licensed spectrum. Soft energy combining schemes provide optimal detection performance by combining the actual sensed information from SUs. For reliable data fusion, these techniques mandate weight estimation for individual SUs in each sensing interval, resulting in high cooperation overhead in terms of time, processing and bandwidth. Alternately, a hard energy combining scheme offers lower cooperation overhead in which only local SU decisions are reported to the fusion center. However, it provides sub-optimal detection performance due to the information loss. In this paper, a Log-Likelihood Ratio (LLR) based cooperative spectrum sensing scheme is proposed in which each SU performs a local LLR based sensing test employing two threshold levels. The local decision and sequentially estimated SNR parameter values (for weight computation) are not reported to the fusion center if the local test result is in-between the two threshold levels. Thereby, cooperation overhead is reduced in proportion to the hard combining techniques; nevertheless simulation results show that the detection performance of the proposed scheme is close to the optimal soft combining techniques.