【作者】 吴寅；

【导师】 刘文波；

【作者基本信息】 南京航空航天大学 ， 测试计量技术及仪器， 2013， 博士

【摘要】 无线传感器网络是新一代的传感器网络，综合了传感器技术、嵌入式计算技术、分布式信息处理技术和通信技术，它的应用和发展将会给人类生产和生活的各个领域带来深远的影响。但是电源供电问题直接影响其性能和生存时间，是限制其实际应用的关键性问题。针对该“瓶颈”，目前众多学者也提出并研究了相当多的能量管理策略，在一定程度上延长了无线传感器网络的生存时间，但是如何从根本上解决供电问题还需进一步研究。本文深入研究了利用外界环境能量为传感器节点供电的自供电无线传感器网络，解决无线传感器网络的自供电问题并对能量进行正确、合理的使用是本文的重点。自供电无线传感节点通过采集环境中的能量来补充供给本身的能耗，可以在很大程度上缓解能量瓶颈并改善网络性能。本文根据能量采集方式的不同分别设计了两种低功耗自供电无线传感节点，详细分析了其组成结构、各子单元能耗模型和系统工作效率，优化了能量采集预测算法和系统能量管理算法，同时本文还分析了自供电无线传感器网络协议层研究的特性，提出了专用的非均匀分簇自适应路由协议，并将协作通信成功应用于网络的簇间通信。具体包含以下几方面内容：1.设计了一种具备完善的自适应能量采集及管理策略的太阳能供电无线传感器节点BFCap，采用双向flyback变换器，实现了最大功率传输跟踪算法，并具有更宽的能量采集范围和冷启动功能。同时还设计了WCSAD算法，可更准确预测下一阶段的可采集太阳能。2.分析并设计了风能供电无线传感器节点的系统能量管理算法。通过解决两个核心问题：最大功率采集和动态电压调节，将风能供电无线传感节点的采能和耗能两个方面有机联系起来，实验证明该算法可达到真正的节点级别的能量最优化管理。3.研究了一种基于强化学习的自供电无线传感器节点能量管理架构，当面对无法预测的环境能量时，该架构不仅可以实时地满足节点维持“永久性”运行的能量需求，而且可以满足自供电无线传感器网络具体任务应用中的服务品质需求。4.设计了一种非均匀分簇的自供电无线传感器网络路由协议，该协议通过在网关中集中式运行大运算量的分簇和路由建立优化算法，而在自供电无线传感节点中分布式运行低功耗的簇首调整和路由更新过程，以尽量提高能量利用率。仿真结果证明该路由可有效平衡网络能量分布并增强通信可靠性。5.针对所设计的分簇型自供电无线传感器网络路由协议，研究了协作通信协议在其上的应用，该协议通过在各个簇内成员节点中分布式运行功率分配及中继选择联合优化算法，来降低簇首间的通信能耗，仿真结果证明该算法适用于所设计的分簇路由并可在保障网络通信性能的前提下进一步提升能量利用率。更多还原

【Abstract】 Wireless sensor network is a new generation of sensor network. This technology is a combinationof sensor technology, embedded computing technology, distributed information processing technologyand communication technology. Its application and development will have far reaching implicationsto various fields of human life and production. However, energy issue directly affects its performanceand survival time, becomes a crucial restriction of practical application. For this "bottlenecks", manyscholars have made a considerable number of energy management strategies to prolong survival timeof wireless sensor networks to some extent, but have not fundamentally solve this energy problem.This article studies the “self-powered” wireless sensor networks which are powered by theenvironment energy. To solve the energy issue and make a correctly and reasonably using of it are thefocus of this article. The self-powered wireless sensor node is supplemented by the harvested energyfrom environment to supply its energy consumption; which makes a great effort to ease the energybottleneck and also improves the network performance. According to the way of energy harvesting,two different design of low-power self-powered nodes are presented, detailed analysis of itscomposition, unit energy consumption of each sub-model and system level efficiency are also shown,simultaneously, optimized energy harvesting prediction algorithm and system energy managementalgorithm also analyzed. In addition, we research the characteristics of self-powered wireless sensornetwork protocol. Dedicated unequal clustering adaptive routing protocol and cooperativecommunication is successfully applied to the network’s inter-cluster and intra-cluster communication.Specifically, it contains the following aspects:1. Design a solar self-powered wireless sensor node BFCap with adaptive energy acquisitionand management strategy; it is made up of a bidirectional flyback converter. It couldachieve maximum power transfer tracking algorithm, and has a wider range of energyharvesting and coldstart function. Also a WCSAD algorithm can more accurately predict thenext stage of collecting solar energy.2. Design a self-powered wireless sensor node energy management architecture based onreinforcement learning, when faced with unpredictable environmental energy. Thearchitecture can not only make the node to meet "permanent" run energy demand real-time,but also can to meet the quality of service requirements in the specific tasks of theself-powered wireless sensor network applications. 3. Analysis and design of wind-powered wireless sensor node system energy managementalgorithm. By addressing two core issues: the maximum power acquisition and dynamicvoltage regulation, the adoption of self-powered wireless sensing system organically linkedand two aspects of energy consumption, the experiments show that the algorithm canachieve true node-level energy optimization management.4. Designed a self-powered wireless sensor network routing protocol based on geneticalgorithm, the protocol by using the centralized operation clustering of large amount ofcomputation and routing in the gateway building algorithm, self-powered nodes distributedrun cluster head of the low-power adjustment and routing update process, in order toimprove as much as possible the network energy utilization.5. Based on the clustering EH-WSN designed the last chapter, a Cooperative communicationprotocol is used here. The protocol uses a distributed application on DF manner, it is a jointoptimization algorithm by the members’ node running power allocation and relay selection,to optimize the energy consumption between cluster heads communications, and it canimprove the network energy utilization ratio effectively.更多还原