【作者】 徐浩漫；

【导师】 宋梅；

【作者基本信息】 北京邮电大学 ， 电子科学与技术， 2013， 硕士

【摘要】 人们对随时随地的进行自由通信的渴望促进了无线网络的迅速发展,对网络系统的容量和高QoS的不断追求给现有的无线通信技术和网络结构带来了巨大的挑战。认知无线电技术因其能够很大程度的提高频谱资源利用率而引起了广泛的关注,频谱动态性是它最大的特性,但是目前对它的研究已远不止最初的动态频谱接入和分配,而是更多的关注产业化应用和商用,将其应用于实际网络和系统,如SON (Self-Organized Network), Small Cell网络、多跳Ad-hoc网络等。本文从理论出发,结合实践,进行了广泛的研究并产生了丰富的研究成果。首先,通过对多信道多用户的资源分配方法进行研究,给出了资源优先(RP, Resource Priority)机制和需求优先(DP, Demand Priority)机制两种通用资源分配机制,前者优先保证最优的资源被合理利用,后者保证优先级最高的用户需求得到满足。本文给出了两种机制下吞吐量、用户间公平性、中断概率、时延以及切换频率等系统性能的比较,并分析用户数量、后备信道的更新时长、平均信道可用时长等因素对于系统性能的影响,并从时间复杂度和存储效率两个方面分析两种机制的优缺点。仿真实验证明RP算法更加适用于信道状态不稳定的情况,其应对信道快速变化的能力较强,而在信道状态相对稳定的情况下,建议采用基于后备信道列表的DP机制,该算法虽然在系统性能上存在劣势,却大大节省了时间开销和存储开销。第二,针对主从用户共存的多业务系统,本文为不具备频谱切换能力的系统提出了种一切实可行的信道分配机制,在保证主从用户通信的同时,提高了系统的吞吐量,通过利用马尔可夫过程的分析和蒙特卡罗仿真结果表明,所提机制能够与具有切换能力的系统相比,在保证中断概率的没有大幅提高的前提下,降低了用户接入的阻塞率最后,针对Overlay共享模式下的认知无线电多跳系统,通过对自物理层至应用层的分析和约束,提出了一种跨层优化分析的方法该办法中以约束路由跳数的系统概率容量最大化为优化目标,并给出了网络接入层的资源分配和调度、网络层的路由选择、传输层连接和流量控制以及应用层QoS指标等约束条件,通过对该问题的分析,提出了启发式的跨层优化算法,并对系统性能和算法性能进行了仿真和分析。更多还原

【Abstract】 The booming of wireless communications is largely furthered by the desire of freely communicating anywhere and anytime. Actually, there is big challenge to the wireless communication technology and structure since the command for system capacity becomes larger and QoS requirement becomes higher. Cognitive Radio (CR) has so big potential to increase spectrum utility that it has been paid close attention to. Researchers take advantage of its spectrum dynamics, and their research works have been far beyond dynamic spectrum access and allocation recently, and they concentrate more on industrial and operational deployment. They apply cognitive technology into practical networks and system, such as Self-Organized Network, Small Cell Network and multi-hop Network, and so on. In the paper, research is done by theory associated with practice and plenty of research results are put forwards.Firstly, two types of channel allocation scheme are proposed after investigation of radio resource allocation methods, that is, Resource Priority Scheme and Demand Priority Scheme. The former one mainly ensures reasonable utility of channel resource and the latter one mainly concentrate on satisfying users’demands. System Performance such as throughput, fairness of users, interruption rate, delay time and spectrum handover frequency is analyzed by theory and simulation. What’s more, the effect to performance caused by number of users, update time of reserved channel list and average available time of channels is investigated, too. Time complexity and storage efficiency are also analyzed to demonstrate the merits and demerits of the two schemes. Numerical results indicate that Resource Priority Scheme is more suitable to channel condition with high dynamic, for it is facilitated with the ability to reply rapid change of channel condition, while in the moderate channel environment, the other scheme with reserved channel list is more appropriate, for its complexity is much lower that the former one, even though there is a little decrease of system performance.Secondly, a kind of practical channel allocation scheme is proposed to solve problems in cognitive systems without spectrum handover ability. The scheme is designed to improve system throughput while ensures the communications of primary users. By the analysis of Markov process and Monte Carlo simulation results, the proposed scheme behaves better than those with spectrum handover ability in service blocking probability while guarantees that interruption rate is not raised largely.Lastly, analysis and cross-layer design is put forward for multiple hops cognitive radio system in Overlay fashion. Through analysis of constraints from physical layer to application layer, cross-layer optimal design is proposed. In the design, optimal object is the maximal router constraint system probable capacity, and the constraints conditions covers resource allocation and scheduling in network access layer, router selection in network layer, connection establishment and flow control in transmission layer and QoS requirement in application layer. Heuristic Algorithm is proposed to solve the complex problem and system performance and algorithm evaluation is put forward to simulate and analyze.更多还原