Enhanced Throughput of Cognitive Radio Networks by Imperfect Spectrum Prediction

Abstract

Spectrum sensing is used to detect spectrum holes and find active primary users while randomly selecting channel for sensing lead to secondary user's low throughput in high traffic cognitive radio networks. Spectrum prediction forecasts future channel states on the basis of historical information. A new frame structure is proposed in this letter for the imperfect spectrum prediction, resulting to select channels for sensing only from the channels predicted to be idle. Simulation results show that secondary user's throughput is significantly enhanced by imperfect spectrum prediction. The impacts of traffic intensity, prediction errors, and channel number on the throughput are also investigated in this study.