LAUSR

Practical cognitive radio networks (CRNs) would have users with different capabilities and access levels to prior information about the primary users (PUs). For instance, some users may have the privilege to access information about the PU network, which can be utilized to increase the detection performance, whereas others may need to use simple detectors due to energy consumptions constraints, etc. This level of heterogeneity must be accommodated to enable coexistence of different secondary users’(SUs) networks. This paper presents performance analysis and comparison of hard and soft combining cooperative spectrum sensing schemes in heterogeneous CRNs. A centralized approach is considered for cooperative spectrum sensing with SUs that may use either energy detector, cyclostationary detector, pilot-based detector, or orthogonal frequency division multiplexing-based detector. For hard combining, each cooperative SU senses the spectrum using one of the available detectors and reports one-bit local decision to the fusion center which then applies $m$-out-of- $K$ rule for the final decision. For soft combining, we derive an optimal soft combining scheme based on the Neyman–Pearson criterion. Then, we drive a suboptimal rule to reduce the complexity of the proposed scheme. For the sake of simplicity, we consider only energy and cyclostationary detectors for soft combining. In addition, we study the impact of the cyclic frequency offset in case of a cyclostationary detector. Finally, the paper is supported with extensive simulation results that demonstrate the performance of proposed schemes in terms of probability of detection and probability of false alarm.