Spectrum Handoff Model Based on Preemptive Queuing Theory in Cognitive Radio Networks
Yang Xiao-Long,Tan Xue-Zhi,Guan Kai
DOI: https://doi.org/10.7498/aps.64.108403
IF: 0.906
2015-01-01
Acta Physica Sinica
Abstract:Cognitive radio can significantly improve spectrum efficiency by temporarily sharing under-utilized licensed frequency with primary users. Its spectrum management framework consists of four parts: spectrum sensing, spectrum decision, spectrum sharing and spectrum handoff. The last part is what we focus on in this paper. Spectrum handoff, which aims at guaranteeing requirement for service of secondary users and shortening time delay produced by interruption from primary users, is an important functionality of cognitive radio networks. For solving the problem of optimizing the extended data delivery time, a spectrum handoff model is proposed based on the preemptive resume priority M/G/m queuing theory. In order to minimize the extended data delivery time, the queuing method with mixed queuing and parallel service is adopted. In this model, each channel has its own high-priority queue and there is only one low-priority queue for all secondary users. The primary and secondary users respectively enter into the high-priority and low-priority queue to establish corresponding primary connections and secondary connections and execute corresponding data transmission. On the above basis, secondary users’ channel usage behaviors are thoroughly analyzed in the cases of multiple secondary users, multiple licensed channels and multiple spectrum handoffs. In this process, when multiple interruptions occur, the secondary user will stay on the current channel and suspend data transmission until primary users finish their data transmission, otherwise the secondary user will switch from the current channel to the predetermined target channel to resume his unfinished data transmission. The target channel is sequentially obtained from the target channel sequence, which is determined by channel parameter estimation algorithm. Based on the analysis of channel usage behaviors for secondary users, the total time delay caused by spectrum handoffs within the whole data transmission process is derived first. The total time delay can be deduced from two scenarios. One is that the target channel is the current channel. For this reason, the total time delay equals transmission time of primary users in high-priority queue. Obviously, the other is that the target channel is not the current channel. Thus, the total time delay equals the sum of transmission times of primary users in high-priority and secondary users ahead in low-priority. In addition, appearance of new primary users should also be considered in the data transmission process. Then, expressions of the extended data delivery time in two different cases (i. e. always-staying strategy and always-changing strategy) are respectively derived. Furthermore, the adaptive spectrum handoff strategy is finally discussed, which is to choose the optimal scheme from always-staying and always-changing strategy when a spectrum handoff happens. Simulation results verify that this model can not only describe handoff behaviors of secondary users more perfectly, but also can make the transmission time delay smaller and make the extended data delivery time shorter than the existing spectrum handoff model. Especially, with the increase of service intensity of primary users, the advantages of the proposed spectrum handoff model are more outstanding. In addition, the allowable secondary user service intensity is improved and the receptive number of secondary user is increased in cognitive radio networks. All in all, the proposed spectrum handoff model improves the performance of spectrum handoff, increases the capacity of cognitive radio networks and optimally realizes spectrum sharing between secondary users and primary users.