Relay selection and power allocation with minimum rate guarantees for cognitive radio systems

This paper investigates joint relay selection (RS) and power allocation (PA) in cognitive radio (CR) systems, in which relay nodes operate in amplify-and-forward (AF) mode. In contrast to the conventional schemes, we take into account the performance of the secondary user (SU) which acts as the relay. A two-step optimization scheme is proposed to maximize system throughput with minimum rate guarantees for the source and the relay while the interference introduced to the primary user (PU) is kept below a specified threshold. The PA optimization in the first step is non-convex, and hence hard to solve in general. To tackle this, we introduce a power proportionality factor at the relay to decompose the non-convex optimization into two levels. Based on dual method and search method at inner level and outer level respectively, a closed-form solution for the optimal PA between the source and the relay is derived. In order to reduce complexity of repeating the PA procedures at all the relay candidates while maintaining reasonable performance, we also develop a suboptimal approach. Furthermore, simulation results and comparisons are presented to illustrate the performance of the proposed scheme.