基干拟生灭过程的无线Ad Hoc网络若干技术研究

发布时间：2018-05-22 15:45

本文选题：Ad + Hoc　；参考：《南京理工大学》2014年博士论文

【摘要】：无线Ad Hoc网络的组网灵活性及抗毁性等特点,使其在军用和民用领域内得到了广泛的应用,但是其剧烈变化的拓扑结构、有限的带宽资源以及共享信道上的竞争冲突给Ad Hoc网络技术的发展带来了挑战,也使得无线Ad Hoc网络的相关技术成为研究的热点。本文利用生灭过程和拟生灭过程对Ad Hoc网络的分簇算法、节点故障维护策略以及MAC层接入协议技术进行了建模和分析,具体体现在以下几个方面：1)建立了多信道接入的分簇网络中簇头节点处理数据请求过程的生灭模型。利用概率分析的方法,引入阻塞概率和信道的平均利用率作为额外的簇头选举因子对分簇Ad Hoc网络中的按需加权自适应分簇算法(Adaptive On-demand Weighted, AOW)进行了改进,并进行了仿真实验,实验结果表明,改进算法后,减少了数据请求阻塞,提高了簇头的平均服务时间,使网络拓扑结构更加稳定。2)建立了无线Ad Hoc网络中按需路由协议路由发现过程中源节点生成子树生灭过程模型(Sub-Tree-BD模型),利用矩阵分析方法进行求解,推导出了生成子树规模变化以及生存时间的概率表达式；基于该模型,建立了Ad Hoc网络节点故障及生成子树维护过程的拟生灭模型,利用概率分析方法,推导了在系统达到稳定状态下,节点故障率、维护过程空闲策略以及系统载荷之间的关系式,对空闲策略对网络性能和网络载荷的影响进行了分析,并通过仿真实验进行了验证,实验结果表明,该模型可优化Ad Hoc网络中的资源分配,为制定维护策略提供参考。3)在对单跳和多跳Ad Hoc网络中MAC层上分布式协调功能(Distributed Coordination Function, DCF)协议的节点数据发送过程分析的基础上,依次建立了单跳网络无限队长拟生灭模型、单跳网络有限队长拟生灭模型、多跳网络无限队长拟生灭模型和多跳网络有限队长拟生灭模型,利用矩阵分析和数学归纳方法对模型进行分析求解,推导了节点缓冲队列无限大时避免网络达到饱和的控制条件,并建立了数据速率的阈值方程；推导了节点缓冲队列为有限时过程的平稳分布,建立了丢包率、平均延迟等网络指标的概率表达式。并通过仿真实验进行了验证,实验结果表明,理论结果与实验数据吻合较好,表明该模型可以描述Ad Hoc网络节点数据发送过程中的竞争退避过程,该模型为Ad Hoc中通过数据速率控制避免网络拥塞提供了一个量化依据,也为特定网络规模下路由协议的选择提供了一种参考。4)在多跳网络有限队长拟生灭模型和二进制指数退避(Binary Exponential Backoff, BEB)算法的基础之上,提出了一种基于拟生灭过程的动态最大重试次数退避算法。引入动态最大重试次数以及节点数据发送速率作为影响因子对BEB算法进行改进,推导出了考虑延迟和丢包率约束的平衡方程,并进行了仿真实验,实验结果表明,改进算法后,与标准退避算法相比,延迟和丢包率显著改善,为网络服务质量控制提供了一种参考。
[Abstract]:The wireless Ad Hoc network has the characteristics of flexibility and destruction resistance, which makes it widely used in military and civil fields. However, the topology structure, limited bandwidth resources and the competition conflict on the shared channel have brought challenges to the development of Ad Hoc network technology and the related technologies of wireless Ad Hoc network. In this paper, the clustering algorithm of Ad Hoc network, node failure maintenance strategy and MAC layer access protocol technology are modeled and analyzed by using the birth and death process and the quasi birth and death process, which are embodied in the following aspects: 1) the birth and death process of cluster head nodes in the cluster network with multi channel access is established. Using the method of probability analysis, the congestion probability and the average utilization rate of the channel are introduced as an additional cluster head election factor to improve the Adaptive On-demand Weighted (AOW) algorithm (Adaptive On-demand Weighted, AOW), and the simulation results are carried out. The experimental results show that the improved algorithm reduces the data after the improved algorithm. Blocking, improving the average service time of the cluster head and making the network topology more stable.2), the source node generation and death process model (Sub-Tree-BD model) of the source node in the routing protocol routing of the wireless Ad Hoc network is established, and the matrix analysis method is used to solve it, and the variation of the generation tree size and the survival are derived. Based on the model, a quasi birth and death model of Ad Hoc network node failure and the maintenance process of subtree is established. By using the probability analysis method, the relationship between the node failure rate, the idle strategy of maintenance process and the system load is derived by using the probability analysis method. The influence of network load is analyzed and verified by simulation experiments. The experimental results show that the model can optimize the allocation of resources in Ad Hoc network and provide a reference.3 for the development of maintenance strategy. The distributed coordination function (Distributed Coordination Function, DCF) protocol on the MAC layer of single hop and multi hop Ad Hoc networks is the node. On the basis of data transmission process analysis, a single hop network infinite captain quasi birth and death model, a single jump network finite captain quasi birth and death model, a multi hop network infinite captain quasi birth and death model and a multi hop network finite captain quasi birth and death model are established in turn, and the model is analyzed and solved by using matrix analysis and mathematical induction, and the node delay is derived. When the impulse queue is infinite, the control condition of the network is avoided and the threshold equation of the data rate is established. The stationary distribution of the node buffer queue as the limited time process is deduced. The probability expression of the network index, such as the packet loss rate and the average delay, is established. The experimental results show that the theoretical results have been verified. The experimental results show that the theoretical results are obtained. It is in good agreement with the experimental data, which shows that the model can describe the competitive backoff process in the node data transmission of Ad Hoc network. This model provides a quantitative basis for avoiding network congestion through data rate control in Ad Hoc, and provides a reference.4 for the selection of routing protocols under specific network scale). On the basis of the quasi birth and extinction model and the Binary Exponential Backoff (BEB) algorithm, a dynamic maximum retrial time backoff algorithm based on the quasi birth and death process is proposed. The dynamic maximum retrial times and the node data transmission rate are introduced as the influence factors to improve the BEB algorithm. The equilibrium equation of delay and packet loss constraint is considered and the simulation experiment is carried out. The experimental results show that, after the improved algorithm, the delay and packet loss rate are significantly improved compared with the standard backoff algorithm, which provides a reference for the quality control of network service.
【学位授予单位】：南京理工大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN929.5

【参考文献】