【作者】 张静；

【导师】 周正；

【作者基本信息】 北京邮电大学 ， 电路与系统， 2011， 博士

【摘要】 资源管理及决策是认知无线网络中受到广泛关注的关键技术,具有广阔的应用前景。本文选题来源于国家863计划等项目,具有重要的理论意义和实际意义。本文在对认知无线网络决策与资源管理的基本原理进行了深入研究的基础上,主要完成了以下具有创新性的研究成果：针对低信噪比环境下单用户频谱感知效率低的问题,提出了一种基于分布式Ad Hoc网络的协作波束形成技术。该技术利用了协作波束形成具有方向性增益的特点,能够有效地抑制噪声,提高了接收信噪比增益,且不需要任何先验信息。仿真表明该方法能够有效提高频谱感知效率。针对认知无线网络频谱分配问题,提出了一种基于层次分析(AHP)算法和图论着色定理相结合的频谱决策分配模型。模型中考虑了频谱可用时间、带宽、延迟等因素与业务特性进行匹配,选出最合适的频谱资源分配给用户。两种算法相结合细化了频谱决策分配机制,既保证了次用户的使用效益又兼顾了网络整体效益,提高了频谱利用率,并能降低次用户的频谱切换次数。针对认知无线网络功率控制问题,提出了一种基于非合作博弈理论的功率控制模型。模型中设计了一种基于信干比(SIR)的代价函数,兼顾了主用户对次用户的干扰因素。仿真结果表明该算法提高了功率利用率,同时还提高了认知无线网络中次用户容量和收敛速度,能够很好地满足认知无线网络环境中对功率快速高效分配的需要。针对认知引擎中多目标优化算法效率低的问题,提出了一种基于进制量子粒子群算法的认知引擎模型。由于算法引入了量子特性,使其具备非线性和不确定性的特点,收敛速度快且精度高,弥补了传统优化算法寻优效率低的缺点。通过OFDM系统验证了算法的高效性,能够很好地在认知引擎中起到多目标优化的作用,达到了对无线资源优化决策的目的。论文最后对全文进行了总结,并对与论文研究方向相关的领域进行了展望。更多还原

【Abstract】 Decision making and Management are key technologies in cognitive radio networks, which are of widespread concern and have broad application prospects. This dissertation is supported by the 863 Project and other funding, and has important theoretical and practical significance.Based on the comprehensive study on the basic principles of resource decision making and management in cognitive radio networks, the main contributions and innovative ideas in the dissertation include:For the problem that single user tend to be low efficient in spectrum sensing in low SNR environments, this paper proposed a collaborative beamforming technique based on the distributed ad hoc network. While this method does not require any prior information, it can effectively suppress noise and improve the received SNR gain by taking advantage of the directional gain of beamforming. Through the analysis based on the energy detector, this method is proved to be effective in enhancing the efficiency of spectrum sensing.For the problem of spectrum allocation in cognitive radio networks, a spectrum decision making and allocation model which incorporates the AHP (Analytic Hierarchy Process) algorithm and the graph coloring theory is proposed. The model considers matching some factors of the spectrum, such as the spectrum available time, bandwidth, and delay, with different service characteristics to allocate the most appropriate spectrum resources to specific users. The combination of the two algorithms can refine the spectrum decision making and allocation mechanism, guarantee the utilization of both secondary users and the entire network, improve the network spectrum efficiency, and reduce the number of spectrum switches of secondary users. For the problem of power control in cognitive radio networks, this paper proposes a power control model based on the non-cooperative game theory. In this model a SIR-based cost function is devised, which takes into account the interference from primary users to secondary users. Simulation results show that the algorithm can improve the power efficiency, increase the capacity of secondary users in cognitive radio networks, and achieve faster convergence speed, well adapting to needs of fast and efficient power allocation in the environments of cognitive radio networks.For resolving the low efficiency of multi-objective optimization algorithms in cognitive engines, a binary quantum particle swarm optimization algorithm is proposed for cognitive engines. Due to the introduction of the quantum properties which can provide nonlinear and uncertain characteristics, the algorithm has faster convergence speed and higher accuracy, thus making up for the low efficiency of traditional optimization algorithms in seeking optimal solutions. The performance of the algorithm is verified through the OFDM system, and the simulation results show that the algorithm can play a favorable effect of multi-objective optimization, achieving the goal of optimizing radio resource decision making.Finally, the content of the whole dissertation is summarized and several valuable research directions are also discussed.更多还原