中文摘要：

认知无线电是在当前频谱资源利用率低下的背景下提出的一种新型无线通信技术．认知无线设备通过自适应频谱感知寻找空闲频谱，实现频谱动态分配和频谱共享，显著提高可用频谱利用率．单信道认知无线系统已有很多研究，但是并行信道系统由于其自身的系统复杂性，研究成果较少．这里利用二状态的Markov链（Markovchain）来描述系统的数据到达特征和并行系统的信道特征，同时利用随机网络演算理论对数据到达过程和信道服务过程进行了形式化建模．通过建立随机到达曲线和随机服务曲线，刻画出数据流的到达特征和并行信道的服务特征，进而得出时延与积压的随机性能边界，对认知无线网络的性能进行评价．最后．通过数值分析与MatIab仿真验证了框架的有效性．

英文摘要：

The cognitive radio （CR） is a newly proposed wireless communication technology to improve the spectrum utilization. Through channel sensing and learning, cognitive radio systems exploit the unused spectrums and occupy it so that the network can allocate the spectrum dynamically and improve the utilization of available spectrum. There are a number of research focusing on the single-channel cognitive radio systems, however the CR system with parallel channels is an unexplored area because of its complexity. In this paper, we conduct performance evaluation of the CR system with parallel channels was conduct by using stochastic network calculus （SNC）. The data arrival process for all prirnary/secondary users and the service process of parallel channels are characterized by Markov modulated deterministic processes （MMDP）. Then we formulate the data arrival process and service process as arrival curve and serviee curve in the SNC theory. The closed form solutions of stochastic performance boundary of the delay and backlog are obtained using the delay bound theorem and backlog theorem in SNC theory. Finally, we make extensive simulations to evaluate the correctness of the proposed model, and conduct performance analysis of the CR system with parallel channels with various parameter settings. The results show that our proposed framework can effectively model and analyze the CR system with parallel channels.