Abstract
A degree of cooperative diversity has been achieved with the introduction of relay nodes in wireless networks. However, such approach suffers from spectral efficiency loss. Therefore, different relay selection techniques have been proposed in the literature to mitigate this loss. In this paper, the problem of relay selection for broadcast cognitive radio networks (CRNs) is investigated. The CRN is composed of a primary transmitter and receiver, a secondary source and several destination nodes. It is assumed that the secondary source broadcasts a signal to N secondary destinations. We implicitly split the destination nodes into two sets: reliable nodes having an signal-to-noise ratio (SNR) larger than a predefined threshold (T) and unreliable nodes with SNR lower than T. The paper proposes several relay selection techniques based on the average SNR, instantaneous SNR, and best direct link. Relay selection is performed by a central node taking into account the interference level at primary receiver and channel state information between nodes of the secondary CRN. In addition, the paper provides the mathematical derivation and analysis of the bit error probability (BEP) of relay selection in broadcast CRNs for both amplify and forward and decode and forward relaying. We propose to select a single relay among reliable nodes that verify interference constraints (interference to primary receiver lower than the interference threshold ) to help the remaining unreliable nodes. The threshold T is optimized numerically to yield the lowest BEP. Our theoretical results are confirmed by simulation.