针对认知无线电网络中认知用户广义传输时间的优化问题,提出了一种基于抢占式续传优先权M/G/m排队理论的频谱切换模型.在该排队模型中,为了最小化认知用户广义传输时间,采用混合排队-并列式服务的排队方式.在此基础上,深入分析多个认知用户、多个授权信道、多次频谱切换条件下认知用户信道使用情况,从而推导出广义传输时间表达式.最后探讨了该模型下自适应频谱切换策略.仿真结果表明,相比于已有的频谱切换模型,该模型不仅能够更加完整地描述认知用户频谱切换行为,而且使得认知用户传输时延更小,广义传输时间更短.此外,认知无线电网络允许的认知用户服务强度增加,能够容纳的认知用户数量增多.因此,该模型提升了认知用户频谱切换的性能,更好地实现了认知用户与授权用户的频谱共享.
Cognitive radio can significantly improve spectrum efficiency by temporarily sharing under-utilized licensed fre- quency 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 interrup- tion 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 trans- mission. 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 ca