Primary User Detection for Cognitive Radio Networks Having Multi-Antennas at Their Base-Stations

Spectrum sensing is a key enabling functionality in a cognitive radio (CR) network. The fusion problem of cooperative spectrum sensing, for a cognitive radio network equipped with multi-antennas at its base-station, is considered in this paper. The base-station equipped with multi-antennas act as a fusion center which can attain not only the soft information (average power) by its multi-antennas but also the soft or hard information from other cognitive radio users (CR users). We first discuss the soft fusion scheme in which soft information from CR users is sent to the base-station, and exhibit the necessity of deeper cooperation. Due to the limitation of control channel bandwidth, CR users have to send hard information to the base-station. To this end, we propose a hybrid fusion scheme in which both soft and hard information can be fused in the base-station. Furthermore, we present an optimal cooperative spectrum sensing scheme where the global detection probability can be maximized according to the Neyman-Pearson criterion. Extensive simulation results validate that, by using the hybrid fusion scheme, significant improvement on sensing performance can be achieved. 