【作者】 唐伟；

【导师】 郭伟；

【作者基本信息】 电子科技大学 ， 通信与信息系统， 2010， 博士

【摘要】 无线传感器网络通常是由一组小体积、低功率且具有无线通信能力的传感器设备所构成的无线多跳网络,有机地结合了分布式信息收集、计算处理及无线通信技术,具有数据中心、面向应用及资源约束等特征,拥有广阔的应用前景,同时也提出了许多挑战。由于无线传感器节点能源的苛刻限制,节能是无线传感器网络中最大的挑战之一,对无线传感器网络设计的各个方面都产生了直接的影响。本文围绕无线传感器网络的节能策略及算法进行了研究,主要研究了单点聚合、多点聚合与多点聚合多径路由策略下最大网络生命期优化路由问题以及空间剖分策略下最小网络总能耗基站位置优化选择问题。针对数据聚合无线传感器网络,延长网络生命期的关键在于采用数据聚合减少网络负载的同时均衡节点能耗,依赖于数据聚合策略的选择。在第二章中研究了单点聚合路由策略。在该策略中,节点选择单个邻居进行数据聚合。提出了以最小能耗生成树为基础结构的数据聚合点优化选择算法,采用遗传算法与梯度下降联合优化的方法对节点能耗进行均衡,提高网络生命期。第三章在第二章的基础上放松了对数据聚合点选择的单一性,考虑了多点聚合路由策略,允许节点将原始数据分发到各相邻节点进行数据聚合。提出了基于最小能耗生成森林的原始数据分配算法,采用线性规划的方法优化节点对原始数据流的分配。应用次梯度方法设计了相应的分布式算法,通过节点间的数据交换与计算实现原始数据流的优化分配。第四章在第三章的基础上进一步放松对聚合数据传输路径的单径约束,考虑了多点聚合多径路由策略,允许聚合数据通过多条路径进行传输。设计了基于线性规划的路由算法获得最大网络生命期,对聚合数据流中的环路进行了消除。提出了聚合数据率约束下的网络生命期的优化路由算法,并对网络聚合数据率与网络生命期之间的权衡关系进行了讨论。第五章研究无线传感器网络中以节能为目标的基站位置优化选择问题。依据网络总能耗目标函数的分段光滑性,提出了基于最短路径树的空间剖分策略对问题进行分治,将空间划分为若干剖分单元,节能路由在各剖分单元中具有不变性。讨论了剖分单元的结构及相邻剖分单元的搜索算法,给出了一维空间中的基站位置优选算法,并为二维空间的情形提出了三种启发式算法以获得性能优良的基站位置。最后,第六章对全文进行总结,归纳了文本研究工作的重点,并对未来的研究方向与热点问题作出了展望。更多还原

【Abstract】 Wireless sensor networks are multi-hop wireless networks, which usually consist of small-sized and low-power sensing devices with wireless communications capabilities. They efficiently combine the technologies in distributed information sensing, processing and wireless communications. Wireless sensor networks are characterised by data centric, application orientation and resource constraints, which lend themselves to countless applications and, at the same time, offer numerous challenges. Due to the stringent energy constraints to which the sensing nodes are subjected, energy efficiency is one the greatest challenges offered by wireless sensor networks, and poses direct impacts on their designs. This dissertation studies the energy reserving strategies and algorithms in wireless sensor networks, which focuses upon the design of the maximum network lifetime routing problems under single-point, multi-point and multi- point-multi-path aggregation routing strageties, and the optimal sink position selection problems under space tessellation strageties that aimed at minimizing the overall network energy consumption.In data aggregated wireless sensor networks, the key to prolong network lifetime is to utilize data aggregation to reduce the network load while balance the energy consumption for the sensors, which depends on the selection of data aggregation strategies. In chapter 2, the single-point aggregation routing strategy is studied. In this strategy, each sensor selects a single aggregator for data aggregation. An optimal aggregator selection algorithm based on the minimum energy cost tree is proposed, which combines the optimization power of genetic algorithm and gradient hill-climbing method. The energy consumption of the sensors is balanced, and the network lifetime is improved.Following chapter 2, chapter 3 relaxes the singularity on the selection of aggregators, and considers the multi-point aggregation routing strategy. In this strategy, sensors are allowed to distribute their raw data to every neighboring sensor for data aggregation. A raw data distribution algorithm based on minimum energy cost forest is presented, and linear programming method is used to optimize the allocation of raw data. Subgradient method is adopted to design the distributed algorithm, and the optimization of the raw data distribution is achieved by inter-node data exchange and computations.Chapter 4 further relaxes the single path constraint for the transmission of aggregated data from chapter 3, and studies the multi-point aggregation multi-path routing strategy. In the strategy, aggregated data are forward via multiple paths. A routing algorithm is designed based on linear programming method, the maximum network lifetime is obtained, and the possible loops in aggregated data flow are eliminated. A routing algorithm is designed to optimize network lifetime under aggregated data rate constraint, and the tradeoffs between network aggregated data rate and network lifetime are discussed.Chapter 5 studies the minimum energy sink position selection problems in wireless sensor networks. According to the piecewise smoothness of the overall network energy consumption objective function, a space tessellation strategy based on shortest path tree is proposed to divide and conquer the problem. In the strategy, the space is divided into tessellation cells, and each cell offers invariability for minimum energy routing. The structure of the tessellation cells and an algorithm for searching neighboring cells are discussed. An optimal sink selection algorithm for 1-D space is given; and three heuristic algorithms are presented for 2-D space, which are able to obtain sink positions with good performance.Finally, chapter 6 summarizes the dissertation, concludes the highlights in the studies, and presents the outlooks of directions and hotspots for future work.更多还原