Ergodic capacity and SER performance analysis of amplify-and-forward cognitive relay networks with partial relay selection

This paper investigates the performance of amplify-and-forward (AF) cognitive relay networks with partial relay selection (PRS) in a spectrum-sharing context. The PRS schemes can be divided into two different types: source-relay PRS and relay-destination PRS, which are respectively called SR-PRS and RD-PRS for short. Different from most of the existing studies which only consider interference power constraint on secondary source and relay nodes, transmit power constraint on secondary relay nodes is further considered in our work. In this condition, we derive the lower bound on symbol error rate and the upper bound on ergodic capacity of the two schemes aforementioned over Rayleigh fading channels, which has never been studied before. According to the analyses, the power constraints and the number of secondary relay nodes are the main factors affecting the system performance. Furthermore, the performances of the two schemes are compared by simulations and the correctness of the analytical results are verified by Monte Carlo simulations.