【作者】 周新莲；

【导师】 吴敏；

【作者基本信息】 中南大学 ， 计算机应用技术， 2010， 博士

【摘要】 无线传感器网络以一种新的信息获取方式与处理模式,将逻辑上的信息世界和客观上的物理世界联系起来,改变人类与自然界的交互方式,实现了物理世界、计算世界以及人类社会的互通。数据收集是无线传感器网络最基本的应用,是各种复杂应用的基础,也是现阶段传感器网络研究的重点。目前开展的数据收集协议研究主要针对静态无线传感器网络来展开的,随着近两年移动无线传感器网络成为学术界的研究热点,原来那些针对静态无线传感器网络提出的数据收集协议应用到移动无线传感器网络时,有些完全不适应,有些会导致协议性能大大减弱。论文目的是针对移动无线传感器网络设计一个节能的数据收集协议。主要从移动无线传感器网络数据收集协议所涉及的网络布署方案、分布式分簇技术、节点休眠调度机制、TDMA时隙划分和sink移动策略几个方面,考虑节点移动性给协议带来的影响,遵循尽可能减少网络的能量消耗的设计准则,达到实现一个节能的分簇数据收集协议的目标。本文的主要工作和创新性成果分为以下四个方面：(1)提出了一种移动节点和固定节点混合布署的移动无线传感器网络模型。具体布署方案为：以网格形式布署m个固定节点覆盖整个监测区域Q,固定节点之间距离为ι(略小于自由空间模型的d0),大量移动传感器节点(n-m个)随机均匀布署在整个监测区域内。(2)提出了一种新的高效节能的分布式分簇算法(ECBF)。算法中移动节点选择最近的固定节点形成簇,由固定节点选择簇内移动节点剩余能量值较高且距离最近的节点作为簇头,这种簇头选择策略既节省节点能量又尽可能避免簇头移动离开本簇。由簇头节点负责收集簇内成员的感知数据,并进行融合处理,由固定节点构成路由骨干进行数据转发,最终递交给sink节点。该分簇方案形成的簇大小一致、位置分布均匀,避免了由移动节点担任骨干路由可能带来的失效。其路由实现简单,可靠性高。ECBF分簇算法弥补了LEACH-M协议中没有考虑节点的能量、簇头和移动性、簇的数目不一致、分布不均匀的一些不足。执行ECBF算法,整个网络的广播消息量复杂度为O(n),整个网络的时间复杂度为O(1),仿真实验表明,ECBF数据收集协议比LEACH-M协议有更长的网络生命期和更好的数据传输成功率。(3)提出了一种避免移动节点位置影响的、满足用户期望覆盖率的、高效节能簇内节点跨层调度算法。算法根据覆盖性分析理论推导出针对监测区域满足用户期望感应覆盖率φ的最小簇内活动节点数目k。在簇内只选择能量较高且距离固定节点较近的k个节点工作(尽可能地避免了簇内节点移动离开本簇),其他大部分节点得到休眠,从而实现高效节能的簇内节点轮换休眠调度。然后采用TDMA时分多路复用技术对全网进行统一的时隙划分,其中簇内时隙划分采用并行方式,簇间时隙划分采用串行方式。该时隙划分方案减少了系统延迟,节省了能量,实现了跨层设计。仿真结果表明采用了跨层节点调度算法实现的数据收集协议EDG降低了系统的数据延迟,节能效果明显。(4)提出了一种不依赖于节点地理位置的基于移动sink的分簇数据收集算法MSBDG。该算法解决了无线传感器网络中多跳路由通信时出现能量空洞的“热点”问题。证明了任意边长为L的正方形监测区域中,sink节点沿2(1/2)L/2轨迹移动时,网络用于数据收集的能量消耗最小。sink沿途以最近的固定节点作为根节点动态构建路由树。簇内移动节点感知的数据经簇头进行数据融合计算,然后将融合后的数据沿路由树反向逐跳转发给sink节点。仿真实验表明,MSBDG在节点的平均能耗和网络的生存时间等方面,其性能超过ECBF, EDG等数据收集协议。更多还原

【Abstract】 Wireless sensor networks are regarded as a new information accessing method and information processing mode, which have linked the logic information world with the objective physical world, changed the way of interaction between human, nature and eventually achieved connectivity among the physical world, the computing world and human society. Data collection, the most basic application of wireless sensor networks, is the basis for a variety of complex applications and also the focus of current research on sensor network. The former study on data collection protocols is mainly based on the static wireless sensor networks, but for the current two years, mobile wireless sensor networks has became the hotspot in academic circle. The original data collection protocol which is aimed at static wireless sensor networks, once it is applied to mobile wireless sensor networks, will greatly reduce its performance.This paper is primarily to design an energy-efficient data collection protocol in mobile wireless sensor networks. Based on these several aspects---the mobile wireless sensor network layout, distributed clustering technology, the node sleep scheduling mechanism, TDMA time slot splitting strategy and sink mobility strategy, it considers the node mobility to the impact of the protocol and follows the design criteria of reducing network energy consumption to achieve the purpose of an energy-efficient data collection protocol.The major work and innovative achievements can be divided into the following four chapters: (1)A hybrid mode of mobile wireless sensor networks deployed a mobile node and a fixed node is proposed.Specific deployment program is:first it covers the entire monitoring area Q with a grid deployed m fixed nodes, then fixes the distance between nodes as 1(slightly less than the free space model d0), finally a large number of mobile sensor nodes (n-m) randomly uniform deployed in the whole monitoring area.(2) A new energy-efficient clustering algorithm (ECBF) is proposed.In the algorithm, a mobile node can choose the closest fixed node to form clusters and the fixed node chooses mobile nodes within the high value of the nearest node in cluster as the cluster head. Therefore, this cluster head node selection strategy not only saves energy as much as possible, but also avoids the cluster head to leave the cluster. The cluster head node is responsible for collecting datum and fusing them. Routing backbone is constructed by the fixed node which is responsible for forwarding data and eventually submitted to the sink node. This clustering strategy makes all the sub-cluster with the same size and distributed evenly, avoiding the potential failure caused by backbone routing. This kind of routing, which can be achieved simply, has high reliability. ECBF clustering algorithm makes up for some drawbacks in LEACH-M protocol, which are not taken into consideration of the energy of the node, cluster head and mobility, the inconsistent number of clusters and uneven distribution. Through implementation of the ECBF algorithm, the amount of broadcast messages throughout the network complexity is O (n) and the entire network time complexity is O (1). Simulation result shows that ECBF data collection protocol, which is better than LEACH-M protocol, has a longer network lifetime and more successful data transmission.(3)A cross-layer scheduling algorithm (EDG) which can avoid the impact on the location of mobile nodes and meet the coverage of user expectations and energy-efficient cluster nodes is proposed.The algorithm, based on the coverage analysis theory, deduces the minimum number of nodes in cluster k by meeting coverageφof the user expectations. In the cluster, it only opts the higher energy and k nodes close to a fixed distance (avoiding the cluster nodes to leave the cluster as much as possible) so that most of the other nodes can be dormant and energy-efficient sleep scheduling rotation in the cluster node can eventually be achieved. Then it takes the whole network into a unified slot by adopting TDMA time-division multiplexing technology, in which slots in cluster are applied to a parallel mode and slots between clusters are applied to a serial mode. Therefore, this kind of slot division can deduce system delay, save the energy and achieve cross-layer design. Finally simulation results show that the adoption of cross-layer scheduling algorithm achieves a data collection protocol EDG, which leads to reduction of data latency and obvious energy-saving effect.(4)An algorithm MSBDG, which does not depend on the location of nodes, is based on mobile sink data collection is proposed.It solves the "hot spot" problem---multi-hop routing in sensor networks brings to energy hole. It also proves that when any square of side length of L is in the monitoring area and sink nodes move along the(?)L/2 trajectories, the energy consumption used by data collection can reach the minimum. Sink node builds the routing tree along the nearest fixed node as root node. At the same time, the sensed data in cluster node carries on data fusion calculation through the cluster head, thus the data fusion is flanked by trees to send to sink node in reverse-by-jump. As a result, simulation shows that in the aspect of average energy consumption and network lifetime, the performance of MSBDG surpasses data collection protocols of ECBF, EDG.更多还原