Admission and power control for cognitive radio networks by sequential geometric programming

In cognitive radio (CR) networks, power control is an effective means to limit the interference caused by the CRs upon the incumbent primary users (PUs) to ensure cohabitation of the two systems. When all CR links can not be active at the same time due to excessive interference, an admission control mechanism is necessary to schedule the CR links. Key to both tasks is accurate knowledge of the CR-to-PU channel gains. However, CRs generally face difficulties in estimating the channel gains very accurately, often due to lack of explicit support from the PU systems. In this work, admission and power control algorithms are developed to account for channel uncertainty through probabilistic interference constraints. Both log-normal shadowing and small-scale fading effects are considered through suitable approximations. The resulting problems can be solved via sequential geometric programming. The admission control is based on solving feasibility problems, whereby CR links violating the interference constraints the most are dropped progressively. The feasible point thus found to initialize the power control iterative solver.