QoS Provisioning Spectrum Decision Algorithm Based on Predictions in Cognitive Radio Networks

In this paper, a QoS provisioning spectrum decision algorithm based on predictions is proposed in cognitive radio networks. By recording and analyzing the historical information, the algorithm can learn intelligently the rule of occupied spectrum by licensed users, and can predict the length of a spectrum hole, with the combination of the use´s QoS request to provide a basis for spectrum decision. The simulation results show that, a good performance is achieved under different distribution of spectrum hole and can reduce the probability of spectrum handover significantly.