无线传感器网络中位置感知的若干关键问题研究

发布时间：2018-03-26 06:37

本文选题：无线传感器网络　切入点：位置感知　出处：《东华大学》2012年博士论文

【摘要】：无线传感器网络集中了微机电系统、感知技术、嵌入式计算技术、信息处理技术和无线通信技术,孕育出一种全新的信息获取和处理模式。位置信息对于无线传感器网络至关重要,只有知道了位置信息,传感器网络感知到的各种上下文信息才变得更有价值。本文针对无线传感器网络的位置感知这一研究热点展开论述,具体内容如下：在分析总结了现有的无线传感器网络节点自身定位算法的基础上,提出了一种高精度节点自定位算法CoTOA。在集中式应用环境中,基于时间到达测距技术,运用约束最优化方法,通过乘子法进行迭代计算,巧妙地将定位误差逐渐降到最小,最终求得未知位置的节点的最优坐标。定位的计算过程由网关sink节点来执行,传感器节点不必承担大量计算。此定位方法的架构设计大大节约了无线传感器网络普通感知节点的计算开销。文中采用软件仿真的方法对定位性能进行了分析,仿真结果显示此定位算法可以获得高精度的定位结果,并且此算法的能量消耗很少,符合无线传感器网络的设计要求。目前的路由研究热点主要集中在保证信息传送成功的情况下尽可能节约无线传感器网络的能量上面。本文从这一关键点入手,提出了一种应用于集中式应用环境下基于位置的能量有效路由算法SEGR。该算法假设传感器节点位置已知,采用集中式无线传感器网络架构,整个网络的路由由网关sink节点来计算、更新并分发。网关sink根据节点的位置信息及剩余能量信息,利用Dijkstra算法计算出最佳路径,然后将每个节点的最佳下一跳信息发给相应节点,开始新一轮的数据传送。该架构将路由计算及发布的能量消耗尽可能地转移到网关节点sink上,很大程度上减轻了普通传感器节点的计算负担和通信负担。从而达到节约传感器节点的能量、延长整个网络生命周期的目的。就本文所重点关注的传感器节点自身定位和路由协议来说,目前这两个领域结合的成果非常少,单一技术的研究目的是为实现一个完整系统打基础的。本文将无线传感器网络节点自定位和无线传感器网络路由协议两项技术结合在一起,提出一个协同设计的模式,在设计初期就让两个技术采用一样的模式、一样的参数、一样的规则等,尤其是在节点自定位和计算基于位置的路由阶段去掉了中间的冗余环节,为无线传感器网络节约了不必要的能量消耗。随后对协同设计的效果进行了软件仿真分析。仿真结果证明,协同设计最终达到了节约能量、减少整合两个系统的工作量、提高效率、进一步延长网络生命周期的目的,协同设计的性能也完全可以满足基本应用的需求,为建立位置感知应用系统打下良好基础。在当今越来越普及的社交活动中,大们经常要出席会议、参观展览等。往往很多地方是第一次去,如果当地的服务人员服务不周到的话,用户需要花费大量精力才能找到准确地点。如果系统能准确预测出用户的下一时刻位置,那么根据当时的上下文环境就可以为用户主动提供导向等服务。本文利用无线传感器网络位置感知技术,提出一种在室内环境下的位置预测算法。该算法将大楼内的房间及走廊等建筑拓扑映射成室内地图,在此基础上,利用机器学习、分类等知识,将用户的历史上下文信息和当前上下文信息作为回归模型中的维进行求解,最后根据判别函数的结果预测用户下一时刻的位置。之后文章举例说明了该预测算法的应用,并对模型的可扩展、可定制的灵活特性进行了分析。本文以传感器网络为基础设施,将传感器网络节点自定位、基于自定位结果的路由策略、室内目标定位、目标位置预测算法等各项关键技术协同起来,提出一个室内位置预测的服务架构。并以医院为应用环境,按传感器网络节点部署、节点自定位、基于位置的路由协议、用户注册、用户服务和用户退出这一整套流程,对该应用模型做了详细介绍。为实现室内位置感知预测应用系统提供强有力的技术支持和理论指导。
[Abstract]:Wireless sensor networks focus on MEMS technology, sensing technology, embedded computing technology, information processing technology and wireless communication technology, creating a new pattern of information gathering and processing. The location information is critical for wireless sensor networks, only know the position information and context information in sensor network to become more valuable. The research focus of the location awareness in wireless sensor network are discussed, the specific contents are as follows:
Summarized based on the analysis of the existing wireless sensor network nodes self localization algorithm, put forward a kind of high precision CoTOA. node self localization algorithm in centralized application environment, arrival time based on the ranging technology, by using the constrained optimization method for iterative calculation by multiplier method, cleverly positioning error gradually reduced to a minimum. The optimal coordinate position of the unknown node can be obtained. The calculation process is carried out by the positioning of the sink gateway node, sensor nodes do not need to undertake a large number of calculation. This architecture design positioning method greatly reduces the computational overhead of wireless sensor networks nodes. Common software simulation method used in this paper to analyze the localization performance, simulation the results show that this algorithm can obtain higher localization accuracy, and the algorithm of the energy consumption of small, with wireless sensor network design to Pray.
Routing the present researches mainly focus on ensuring that information transfer the case of success as much as possible to save energy of wireless sensor network. This paper from the above this critical point, the paper presents a design of centralized application environment based on SEGR. energy efficient routing algorithm based on the assumption that the position of the sensor node position is known, the centralized wireless sensor network the whole network architecture, routing is calculated by sink gateway node, update and distribution. According to the sink gateway node location information and residual energy information, using Dijkstra algorithm to calculate the optimal path, each node of the best next hop information to the corresponding node and then start a new round of data transmission. The architecture will be calculated and release the routing energy consumption as much as possible to transfer to the gateway node sink, greatly reducing the calculation of ordinary sensor node negative Burdens and communication burdens to save the energy of sensor nodes and prolong the entire life cycle of the network.
This paper focuses on the sensor node localization and routing protocol, the combination of the outcomes of these two areas are very small, single purpose technology is to lay the foundation for the realization of a complete system. The wireless sensor network node self localization and wireless sensor network routing protocol two technologies together and put forward a collaborative design model, at the beginning of the design to two by the same pattern, the same parameters, the same rules, especially in the node self localization and route calculation stage based on the location of removing the redundancy link, for wireless sensor networks to save unnecessary energy consumption. Then the simulation analysis on the effect of collaborative design. The simulation results prove that the collaborative design eventually save energy, reduce the workload of the integration of the two systems, improve the efficiency in The purpose of extending the network life cycle is one step. The performance of collaborative design can also meet the needs of basic applications, and lay a good foundation for the establishment of location aware application system.
In today's increasingly popular social activities, people often have to attend the meeting to visit the exhibition. Many places are often the first to go, if not good local service personnel service, users need to spend a lot of effort to find the exact place. If the system can accurately predict the user's next time, then according to the context that can provide guidance for the user initiative service. This paper uses the sensing technology of wireless sensor network location, a location in the indoor environment prediction algorithm. This algorithm will be in the building room and corridor building topological mapping into indoor map, on this basis, the use of machine learning, classification etc. the history of knowledge, the user context information and the context information as a regression model in dimension calculation, according to the results of the discriminant function to predict the user next time The article illustrates the application of the prediction algorithm, and analyzes the extensible and customizable flexibility of the model.
This paper based on sensor network facilities, the sensor network node self localization, routing strategy based on the results of indoor positioning, target location, key technology of target position prediction algorithm and other synergies, put forward a prediction of indoor location services architecture. And in the hospital environment for the application, according to the sensor network node deployment node location, location based routing protocol, user registration, user service and users out of the whole process, the application of the model in detail. In order to realize the prediction of indoor location aware to provide strong technical support and theoretical guidance for the application system.

【学位授予单位】：东华大学
【学位级别】：博士
【学位授予年份】：2012
【分类号】：TN929.5;TP212.9

【参考文献】