Performance of AF Relay Selection With Outdated Channel Estimates in Spectrum-Sharing Systems

In this letter, we investigate the performance of amplify-and-forward partial relay selection in underlay cognitive radio systems with delayed channel estimates, where a direct secondary link is considered. We derive the exact outage probability using a novel mathematical approach based on the moment generating function. Unlike the previous works, our analysis shows that for varying interference-to-noise ratio (INR), i.e., γ̅_I = εγ̅, where 0 ≤ ε ≤ 1, the diversity gain is a function of ε regardless of the nature of the channel state information (CSI). Under the varying INR assumption, it is found that the diversity gain is improved for both outdated and perfect CSIs when ε → 0. The validity of our analysis is verified by computer simulations.