【作者】 胡青松；

【导师】 焦李成； 张申；

【作者基本信息】 中国矿业大学 ， 通信与信息系统， 2011， 博士

【摘要】 煤矿隧道中的无线信道是受多种环境参数影响的动态模型。现有无线通信系统的设备参数由于无法随煤矿信道模型自适应的调整,因此通信效果不佳,而全认知无线电是解决该问题的有效途径。本文的目的就是要将全认知无线电网络能够感知环境参数并对设备进行自动重配的特点用于煤矿隧道通信,解决复杂多变的矿井环境下的自适应通信问题。本文设计了一个通用的煤矿认知无线电网络模型,研究了在该网络中的节点部署方式,给出了认知信息的处理与应用框架。该网络模型通过区域性认知无线电网络完成环境信息的感知和收集,由认知语言规范来支持对用户需求的理解。节点在巷道中部署成不规则的链状形式,而在井底车场等区域将各节点聚集成簇,借鉴无线传感器网络中的分簇路由算法进行数据的转发。数据集被抽取为本体信息库和元数据库,通过数据翻译、业务流程描述、数据路由系统等功能模块,将内容按需路由给特定的应用系统。在矿井这种时变、动态的环境中,许多通信细节事先无从知晓,因此设计了一个基于改进蚁群算法的路由协议解决启发式路径探寻问题。另外,由于矿井隧道是一个有限空间,数据没有必要全向广播,因此研究限制节点的广播角度的方法,以节省路由开销。该协议采用前向蚂蚁进行路径探测,反向蚂蚁更新各节点的路由信息,出错蚂蚁对路由进行报告和维护。本算法综合考虑了端到端时延和节点可用SOP(频谱机会)信息,避免了负载不均的问题。发现了分簇路由算法中的零簇头和簇头失衡现象。零簇头问题会导致虚假轮的出现,浪费节点的能量,影响算法分析的真实性。簇头失衡问题会使得各节点的能量消耗严重不均,致使部分节点快速死亡。在分析其成因的基础上,提出了自适应修改阶段轮数、调节因子和混合法三种解决方案。将分簇思想用于煤矿认知无线电网络的路由协议设计中,提出了两个分簇路由算法,即层次型分簇路由算法LCRA和基于Voronoi图的能量均衡分簇路由算法EBCRV。LCRA的特点是根据网络拓扑特征,对网络进行结构上的划分。而EBCRV不但考虑了簇头选举的随机性,而且照顾到了各轮中簇头数目的均衡性。同时,将节点剩余能量纳入簇头选举指标,均衡了簇头和普通节点的能量消耗。该论文有图60幅,表6个,参考文献132篇。更多还原

【Abstract】 The wireless communication channel of colliery tunnel is a dynamic model affected by many environmental parameters. Existing wireless communication systems can’t change their device parameters automatically with the change of colliery channel model, so have bad performances, especially in tragedy cases. The full cognitive radio may be a good solution to this problem. So, we want to use the full cognitive ratio to resolve the adaptive communication problems in mine, making communication devices reconfigure themselves through the circumstance cognition.A universal cognitive radio networks model for colliery and an integrated framework for node deployment and cognitive information processing and application were brought forward in this dissertation. The CRN model senses and collects environmental information through regional cognitive radio networks, receives and understands the users’requirements. The cognitive nodes form unregularly chains in laneway and clusters in larger area whose data can be forwarded using clustering routing algorithms. Datasets are extracted as ontology information base and meta-data base, and their contents are routed to specific applications through data interpretation module, work flow description module and data routing system.Because of the changing and time-varying communication circumstance, many communication details can’t be foreseen by designer. For this reason, we designed a routing protocol based on extended ant colony algorithm to find data path heuristically. Moreover, the mine tunnel is a limited space and need not to broadcast data in every direction, so the broadcast angle can be restricted in a flood scope to decrease routing cost. The protocol explores path through forward ants, updates routing information of intermediate nodes through backward ants, reports and maintains existing path using error ants. This protocol considers both end-to-end delay and nodes’available SOP information, so can avoid the flow unbalance problem. The zero cluster head and cluster head unbalance problems were discovered in clustering routing algorithms. Zero cluster head problem will lead to dummy rounds which waste the node energy and affect the authenticity of algorithm analysis. Cluster unbalance problem will lead to inequality of energy consumption between nodes, and makes some nodes die quickly. After analyzing the arisen reasons of these two problems, we advanced three solutions, including adaptive round number modification of an epoch, adjusting factor and compound method. Based on the research of clustering principles, we put forward two clustering routing algorithms for colliery cognitive radio networks, one is layered clustering routing algorithm (LCRA), the other is energy balanced clustering routing based on Voronoi graph (EBCRV). The LCRA divides the network into some sub regions based on the topology characteristics. EBCRV considers not only the randomicity of head erection, but also the balance of head number of each round. Besides, the EBCRV take the residual energy of node as one of head erection metrics, making the energy consumption more uniform.There are 60 figures, 6 tables and 132 references in this dissertation.更多还原