【作者】 刘志；

【导师】 裘正定；

【作者基本信息】 北京交通大学 ， 信号与信息处理， 2011， 博士

【摘要】 能量问题是制约无线传感器网络(Wireless Sensor Networks, WSN)大规模推广应用的一个关键因素,是WSN中的研究热点之一。提高能量效率,一方面需要降低能耗总量,另一方面需要均衡能耗。论文这两个角度出发,对WSN中的能量高效覆盖控制和路由问题进行了研究,主要工作和创新如下：对于WSN的覆盖控制问题,论文从活动节点拓扑分布的角度对现有算法进行了研究,提出一种基于准格型策略的活动节点选择算法。设计了分布式的格点坐标求解方法,通过退避式竞争选择节点,减少节点密集分布带来的访问冲突及通信能耗；针对覆盖漏洞问题,设计了邻近节点排斥技术用于补充激活,使活动节点相互远离；针对能耗均衡问题,设计了原点坐标平移主导的虚拟网格平移方案,实现活动节点的周期轮换；论文还讨论了节点分布强度对不规则度及格边长的影响。仿真表明,该算法能够降低活动节点数量,延长网络的生命周期。针对大规模WSN网络由于多对一传输特性而造成的能耗不均衡问题,提出了一种基于分环多跳的能耗均衡分簇路由算法,并着重研究了簇头数量与能耗均衡的关系。算法从最小化第一环总能耗、均衡不同环簇头能耗的角度出发,推导了各环的最优簇头数量,为不均匀分簇过程提供理论指导；进一步,算法设计了-种受控簇头竞争方案,限制簇头分布的随机性。仿真表明,该算法能够更好地均衡能量消耗。WSN数据收集问题的关键是实现网内数据处理过程与路由过程的有效结合。论文研究了贝叶斯压缩感知理论在这一问题中的应用,重点讨论了具有路由结构的稀疏投影矢量的构造算法。由于该矢量的构造是NP难的,将其划分为三个子问题求解：(1)目标节点集选择；(2)路由构建；(3)投影系数计算。对于子问题(1),提出了两个指标指导目标节点的选择：最大广义特征值对应特征矢量的主分量、节点系数总能量。对于子问题(2),建立了问题的形式化表述,并将环形路由问题转化为节点分离的最小代价双组播树问题进行求解。最后,根据前两步确定的非零系数位置,通过最大化微分熵的减少量求解子问题(3)。仿真部分对算法的重建性能、能量消耗、复杂度等指标进行了评估,证实了本算法能够降低计算复杂度和能量消耗,改善重建性能。最后,针对一类特殊的传感器网络—延迟容忍传感器网络(Delay Tolerant WSN, DTWSN)中的能量高效路由问题进行了研究。在优化目标上,论文提出使网络在尽量满足一定服务质量的情况下,以最小化传输代价作为优化目标。在算法设计上,提出了一种根据单个报文的传输状态动态调整路由策略的自适应路由算法。算法根据点的传输能力分配复制指数,使传输能力强的节点拥有更多的报文复制机会；以复制指数比重为指标进行队列管理,减小丢包对传输率的影响；在传输调度方面,算法定义了报文的剩余行进速度,优先调度急需传输的报文。仿真表明,该算法能够在保持一定的传输率、增大端到端延迟的情况下,降低传输代价(额外开销)。更多还原

【Abstract】 Wireless Sensor Networks (WSN), usually battery-powered, thus extremely energy-constrained, present energy efficiency as the main challenge. To improve energy efficiency, one strategy is to minimize the total energy consumption of the whole network; the other one is to balance the energy consumption among nodes. In this dissertation, problem of coverage control and routing in WSN is studied, and four algorithms to improve energy efficiency are proposed. Main contributions of this dissertation are listed as follows:(1) To improve scalability and node’s coverage efficiency, we propose to build a quasi-grid network from randomly distributed sensor nodes by constructing a virtual global grid structure to control the activation process. In the algorithm, a distributed scheme is presented to determine the position of closest vertex for each node, and a neighboring repulsion technique scheme is designed to guide the supplementary activation process. To balance energy consumption, an original point initiated grid shift scheme is also designed. In addition, to tolerant some degree of irregularity, a method to calculate the grid side length is proposed. Simulation results show that the proposed algorithm can reduce the number of active nodes and prolong the coverage lifetime of WSN.(2) To overcome the well-known hot spot problem and simplify routing construction in large area, a new ring based multi-hop clustering routing algorithm is proposed. The whole area is divided into concentric rings and nodes in each ring are arranged into a number of clusters. To overcome energy hot spot problem, optimal number of clusters is derived through balancing energy consumption of cluster heads in all rings. In addition, to confine the distribution of cluster heads in each ring, a new cluster head competition scheme based on sector and grid division is designed. Simulation results show that the proposed algorithm can balance energy consumption among nodes and prolong the network lifetime.(3) The problem of energy efficient data collection is studied, and a new adaptive data collection algorithm based on Bayesian Compressive Sensing (BCS) is proposed. The problem of building a sparse projection vector containing routing structure is proved to be NP-hard. To find a suboptimal solution, it is divided into three sub-problems:target nodes selection, routing construction, and projection coefficients determination. For target nodes selection problem, two metrics are proposed:principal components of the eigenvector with largest eigenvalue, and nodes with the least total coefficients energy. For routing construction problem, a formal description of the problem is formulated, and it is converted into a problem of building double node disjoint multicast trees. Finally, projection coefficients are obtained by maximizing the reduction of differential entropy. Simulation results show that the proposed algorithm can reduce computation complexity, and improve reconstruction quality.(4) Routing problem in a special kind of WSN -- Delay Tolerant WSN (DTWSN) is studied. Firstly, due to extremely energy limitation, we declare that the optimization object for routing in DTWSN should be to minimize the transmission cost, under the constraints of QoS requirements. Secondly, to save energy consumption, based on the progress status of a bundle, an adaptive routing algorithm is proposed, which selects duplication strategy and forwarding strategy dynamically. In addition, mechanisms for queue management and transmission scheduling are designed. Simulation results show that the proposed algorithm can reduce the transmission cost and keep an acceptable delivery ratio, at the cost of increasing in transmission delay.更多还原