A SINR Guaranteed Power and Spectrum Allocation Algorithm for Cognitive Networks

Abstract

Cognitive network is now regarded as a promising approach to solve the spectrum underutilization problem. To protect the primary user, interference temperature has been promoted as a threshold under which the secondary user must control its transmitting power. As the traditional contention graph based model can't actually depict this interference character, we proposed a geometric SINR model which can jointly solve the interference temperature requirement and optimal spectrum assignment problem. Differing from other research works, this paper considers both primary and secondary users' SINR requirement. As finding the optimal results of power control and spectrum allocation is proved to be NP complete, we proposed a novel heuristic power and spectrum allocation algorithm (HPSA) to solve this problem efficiently. Based on the HPSA algorithm's result, an intensive iterative power and spectrum allocation (IPSA) algorithm is designed to find a more excellent result. By detailed simulation, the results demonstrated that this algorithm can quickly be converged within 10 iterations. IPSA also has the average performance improvement about 60% comparing to the maximum coloring scheme.