【作者】 瞿慧；

【导师】 沈连丰；

【作者基本信息】 东南大学 ， 通信与信息系统， 2005， 硕士

【摘要】 无线传感器网络通过一组集成有传感器、数据处理单元和通信模块的结点,通过无线通信技术,以Ad-Hoc方式组网,其目的是通过网络结点间的协作,采集、处理和传输、发布网络覆盖区域中感知对象的信息。随着微机电系统(MEMS)、超大规模集成电路技术(VLSI)和无线通信技术的飞速发展,它的应用空间日趋广阔,遍及军事、民用、科研等领域;但由于网络结点自身固有通信能力限制、能量限制、计算及存储能力限制,对无线传感器网络的研究具有很大的挑战性和宽广的空间。以美国为首的许多发达国家都已经把发展无线传感器网络的研究作为资助重点,但我国对此方面的研究尚在起步阶段;为此,作者详细考察了目前对无线传感器网络的应用和研究现状。无线大规模传感器网络由于在信息质量、网络健壮性、以及网络自适应性等方面的突出优势,得到更为广泛的应用;但结点的微型化不可避免的带来结点可用能量的更严峻的限制,因此,延长网络的生命期已经成为进一步推广大规模无线传感器网络应用的关键。延长大规模无线传感器网络生命期的研究大体可以分为两个方向,其一是研究如何减小结点自身和物理层链路的功耗,而现在更受重视的方向则是研究网络结点之间的协作策略,以最节能的方式完成网络的通信任务,这也是本学位论文的研究着眼点。作者首先给出了对无线传感器网络生命期的准确定义,以及网络的功耗模型,接着提出并论证了网络结点的最佳中继选择策略。在此基础上,作者以单信息源无线传感器网络为研究对象,分析了多种具有代表性的信息源分布与活动规律下,使用最佳协作策略所能实现的最大的网络生命期,即网络生命期的理论上界。随后,作者介绍了在具有中心控制器,或结点具有全局知识,能够进行分布式的中心控制的网络中,通过网络流控制来设计最大生命期的无线传感器网络的方法,论证了它的可行性和优势,并给出了针对不同网络的最优化目标函数和约束条件。考虑到无线传感器网络自身的特点,作者提出了网络可行的协作策略应当具有的特性,即所有网络结点地位等同,都只具备对网络的局部知识,通过分布式的操作实现网络功能;协作策略本质上无环路;协作策略引入能量敏感参数,具备对能量消耗及结点自身能量的敏感性。为了介绍能量敏感的分布式协作策略,作者首先给出了策略对网络的三个假定,即网络具有高密度、结点具备相对于基站的位置知识、结点具备邻居的位置及能量知识;并提出和论证了使用TOA+半-GPS方法进行网络结点定位的有效性和优越性。在此基础上,作者层层深入地介绍了能量敏感的分布式协作策略的基本规则,并通过大量仿真的统计结果论证了该协作策略的可行性,以及该协作策略对提高大规模无线传感器网络生命期的有效性。更多还原

【Abstract】 Wireless sensor networks are constructed by nodes equipped with sensor, data processing unit and wireless communication unit in an ad-hoc mode, and their main function is to collect, process, transmit and broadcast activities of the objects in the observed area by the collaboration of all the network nodes. With the development of Micro-Electro-Mechanical systems(MEMS), Very-Large-Scale-Integration systems(VLSI), and wireless communication techniques, the application of wireless sensor networks has appeared in many fields, including military, civil, scientific, etc. However, due to their compact form factors, network nodes have limited communication, computation and storage capability, and are always severely energy constrained, which makes the study of wireless sensor networks both challenging and promising. Many advanced countries have put great efforts into the study of wireless sensor networks, while this research is still underway in our country. Therefore, the author explored the status of application and research of wireless sensor networks in detail.Large-scale wireless sensor networks composed of thousands of nodes have much broader application because of their superior information quality, better stability and higher adaptability. However, their tiny nodes have much severer energy constraint, and therefore, extending the lifetime of these networks has become pivotal for extending their application. Researches for extending the lifetime of wireless sensor networks fall into two categories. One trend is to consider the design of the power supply system as well as the physical layer of the nodes, while the other trend focuses on the collaborative strategies of the network. The latter has been put more emphasis nowadays, and the research of this thesis goes along this way.The author first presented a detailed definition of sensor networks’lifetime and power consumption model, and then explored the best routine for selecting relay nodes in a multi-hop network. Based on this result, the author focused on the wireless sensor networks with a single information source, and analyzed the theoretical upperbound of the lifetime that could be realized by using the best collaborative strategies under several representative source distributions and activities. Afterwards, the author introduced the technique of maximizing network lifetime with flow control methods for the networks with a central control part or distributed central control function, proved its feasibility and advantages, and listed the detailed object functions and constraints for some representative networks.Considering the specific characteristics of wireless sensor networks, the author pointed out the necessary properties for a feasible collaborative strategy, that is, network functions are carried out by nodes only with local knowledge in a distributed way, the strategy is essentially loop-free, the strategy has some energy-related parameters and is sensitive to the energy consumption as well as the energy level of the nodes. Afterwards, the author introduced and proved the reasonability of three assumptions, namely, the network has high-density, each node is aware of its position according to the sink, and each node keeps track of its neighbors’ positions and current energy levels. Based on these assumptions, the rules of our adaptive distributed collaborative strategy are presented thoroughly. Finally, this strategy’s feasibility and effectiveness in extending the lifetimeof large-scale wireless sensor networks are shown by simulations.更多还原