【作者】 桂丽；

【导师】 邹仕洪；

【作者基本信息】 北京邮电大学 ， 计算机科学与技术， 2013， 博士

【摘要】 随着无线通信技术的飞速发展,迅猛增长的无线频谱资源需求与有限的频谱资源供给矛盾日益尖锐。研究报告的数据表明,目前传统基于授权的固定频谱分配机制使得无线频谱资源无法得到高效利用。认知无线电技术采用动态频谱接入技术,动态发现并利用可用频谱资源,对提高频谱资源利用率起到推动作用。认知无线自组织网络将认知无线电技术延伸至多跳无线自组织网络场景,改变了认知无线网络依赖于中心控制的组网形式,具备传输范围广和结构组织灵活的优点,在时、空、频三个维度充分挖掘并利用可用频谱资源。然而,目前认知无线自组织网络仍面临诸多挑战：首先,无线资源使用的无序竞争使得网络容量受到制约；其次,频谱可用性的动态变化导致无线链路的不稳定,从而使得基于传统网络协议的端到端传输不稳定且服务质量低下；此外,多跳路由及其数据传输由于路径失效进而引发丢包、拥塞等问题。为了解决上述问题,本论文在对认知无线自组织网络功能架构进行分析与梳理的基础上,分别针对链路层、网络层和传输层中功率分配机制、路由机制以及流量分配机制展开研究,旨在提高认知无线自组织网络业务传输的端到端服务质量。本论文的研究内容及主要贡献概括为如下三个方面：针对认知无线自组织网络环境下物理及链路层功率控制所面临次用户对主用户以及次用户之间存在的干扰导致资源利用率低的问题,本论文提出了基于干扰温度限制的分布式功率分配机制,旨在保证主用户正常传输的前提下,最大化次用户网络的系统容量。首先,该机制通过设置干扰测量节点,有效控制次用户对主用户的干扰低于干扰温度限制；其次,次用户基于斯坦伯格博弈理论根据主用户的功率选择通过迭代式求解最优功率控制方案；最后,仿真结果验证了所提机制最优解求解的收敛性,通过与平均分配机制和按比例分配机制之间对比,本论文所提出功率控制机制的次用户系统容量较其他两种分配方案更优并能有效达到功率控制的目的。针对多跳认知无线自组织网络中网络层路由稳定性低以及服务质量低下的问题,本论文提出一种基于三维冲突图模型的多路径路由机制。该机制基于三维冲突图模型对次用户之间干扰和资源利用冲突关系进行描述,并基于冲突图的最大独立集选择多条互不冲突的路径用于数据传输。考虑到认知无线自组织网络中节点能量的有限性,如果始终采用所有路径发送数据,而忽略路径上节点传输信道增益和可用性的变化,将造成节点能量无法用于高效的数据传输,进而导致节点能量和频谱资源的浪费。本论文在认知无线自组织网络多路径路由场景下提出一种最优路由选择方法以提升多路径路由性能。该机制采用马尔可夫链对信道增益、信道可用性以及节点能量变化进行模拟,并采用随机优化方法Restless Bandit理论对问题建模,以传输时延和节点能量消耗作为联合优化目标。仿真结果显示,通过与随机选择机制和贪婪选择机制进行对比,本论文所提机制在有限时间内能在最小化传输时延和最小化节点能量消耗之间达到良好折衷,使得认知无线自组织网络的多路径路由得到高效利用。针对认知无线自组织网络多路径传输受到频谱可用性和业务到达率双重动态影响且缺乏自适应流量分配,进而引发端到端时延和丢包率增加的问题,本论文提出一种认知无线自组织网络基于多路径传输的动态流量分配机制。该机制根据各路径的处理能力,动态分配业务数据流,以达到最优化端到端时延性能的目标。该机制基于排队论对多路径传输下数据到达率、节点处理能力和动态频谱可用率建模,并分析上述动态因素与多路径传输端到端时延的关系,进而将问题建模为最小化各路径最大时延的优化问题。本论文通过对优化问题的迭代式求解,获得多路径传输下的最优流量分配方案。最终,通过仿真验证本论文所提算法的收敛性,通过与平均分配和比例分配机制进行性能对比,验证本论文所提流量分配机制能有效提升认知无线自组织网络的端到端时延性能。最后,论文针对上述认知无线自组织网络相关研究成果进行了全面总结,并针对相关研究方向提出了研究展望。更多还原

【Abstract】 With the rapid development of wireless communication technology, the contradictions between increasing wireless spectrum resource needs and limited spectrum resource supply are becoming increasingly acute. The data of research report shows that, the traditional licensed spectrum allocation policy decreases the efficiency of wireless spectrum resource utilization. The cognitive radio technology adopts dynamical spectrum access technology and dynamically discovers the spectrum opportunity which is impulsive for increasing the spectrum utilization efficiency. The Cognitive Radio Ad Hoc Networks (CRAHNs) extend the cognitive radio technology to the multi-hop wireless ad hoc networks scenario, which changes the centralized network structure of cognitive radio networks, and the large range of transmission and flexible network structure make the spectrum resource discovering on the time, space and frequency domain more sufficient.However, the CRAHNs face lots of challenges nowadays:firstly, the disordered competition for the wireless resources decrease the network capacity; secondly, due to the dynamical spectrum availability, the wireless link becomes unstable which decreases the end-to-end stability and quality of service by using the traditional network protocol; besides that, the break routes cause packet loss and congestion problem during the multi-hop routing and data transmission. To solve the problem described above, this dissertation investigates the architecture of CRAHNs which is the research basis for the dissertation. The dissertation studies the power allocation schemes, routing schemes and traffic assignment scheme on the link layer, network layer and transport layer, in order to increase the end-to-end transmission quality of service in CRAHNs. The dissertation content and contributions are summaried as follows:Considering the interference which exists among secondary users and between the primary users and secondary users on the physical and link layer in CRAHNs, which decreases the resource utilization efficiency, a distributed power allocation scheme, which is based on the interference temperature limit, is proposed in this dissertation to guarantee the transmission quality of primary users and maximize the capacity of secondary user networks. Firstly, the interference monitor nodes are set up in this scheme to ensure that the interference to the primary users is below the interference temperature limit. Secondly, based on the Stackelberg game, the secondary users allocate the power after the primary users and get the optimal power control schemes by iterative solution. Finally, the simulation verifies the convergence property of the proposed scheme. Furthermore, by comparing with the equally allocation and channel gain based ratio allocation, the proposed scheme outperforms other schemes in capacity.Considering the low stability and poor end-to-end quality of service on the network layer in multi-hop CRAHNs, this dissertation proposed a3-dimensional conflict graph based multi-path routing scheme. The multi-path routing scheme characterizes the interference and resource utilization conflicts among secondary users by3-dimensional conflict graph and performs the conflict-free routes selection to transmit data based on the maximum independent set. Considering the limited energy of nodes in CRAHNs, if the changes of channel gain and the availability are neglected in multi-path routing, the energy of node can not be efficient utilized in data transmission which incurs the waste of node energy and spectrum resource. In order to optimize the efficiency of multi-path routing in CRAHNs, an optimal routes selection scheme is proposed in this dissertation. The Markovian chain is used to model the state transformation of channel gain, channel availability and energy level of nodes. The problem is modeled based on the Restless Bandit theory, which is a stochastic process optimization problem, to jointly optimize the transmission delay and energy cost. The simulation results show that the scheme proposed in this dissertation outperforms the random selection scheme and greedy selection scheme in joint optimal utility function of transmission delay and energy cost, and increases the efficiency of routing in CRAHNs.Considering the dynamical factors of spectrum availability and service arrival rate in multi-path transmission CRAHNs, the end-to-end delay and packet loss ratio are enhanced due to the lack of adaptive traffic assignment. A dynamical traffic assignment scheme for multi-path transmission in CRAHNs is proposed in this dissertation. The scheme assigns the traffic to the paths according to the processing ability to attain optimized end-to-end delay. The queue theory is used to model the packet arrival rate, processing ability and spectrum availability of secondary users which affects the end-to-end delay of multi-path transmission. The problem is formulated as a min-max problem which minimizes the maximum delay in multi-path transmission, and gets the optimal traffic assignment proposal by iterative solution. The simulation verifies the convergence property of this scheme. Moreover, the proposed scheme is compared with the equal and ratio assignment scheme which outperforms other schemes in end-to-end delay.In the end, the research is summarized in this dissertation, and the related directions are pointed out in CRAHNs.更多还原