无线传感器网络定位及覆盖技术研究

发布时间：2018-06-10 11:05

本文选题：无线传感器网络 + 节点定位　；参考：《南京理工大学》2016年博士论文

【摘要】：无线传感器网络(Wireless Sensor Networks,WSNs)是随着半导体技术、微系统技术、通信技术等技术的发展而产生和迅速发展起来的。无线传感器网络由各类集成化的微型传感器节点协同感知、采集和处理网络覆盖的地理区域中感知对象的数据,通过嵌入式系统对数据进行处理,并通过随机自组织无线通信网络将这些数据传送给基站,最后通过互连网或卫星网络到达管理节点。无线传感器网络目前广泛地应用于国防军事、国家安全、环境监测和医疗卫生等领域。而在这些应用研究中,节点定位和网络覆盖是无线传感器网络应用的两个研究热点。节点的位置信息对于实现无线传感器网络众多应用起到至关重要的作用;而网络覆盖则决定了无线传感器网络所能提供的服务范围,也在很大程度上影响了网络的成本和各种具体应用的性能。本文针对无线传感器网络中的节点定位与网络覆盖技术进行了比较深入的研究与探讨。在节点定位方面,首先分析改进了传统DV-Hop定位算法,然后以核方法为研究手段,结合主流学习算法,提出了两种新型定位算法;在网络覆盖方面,主要针对静态覆盖中基于虚拟势场的覆盖算法做了研究,提出了基于静电场理论的移动传感器网络部署算法。本文的主要研究内容以及创新点如下:1.提炼出了 DV-Hop算法产生误差的几个原因,在此基础上提出使用粗定位到精确定位的递增式算法,并且在三边测量法计算阶段引入共线度概念,选择定位质量好的单元进行计算,从而使定位算法得到了较高的定位精度;2.借助移动信标节点,使其按照规划的路径移动,从而产生若干个虚拟信标节点,这样可以有效减少真实信标节点的数量。同时将由这些虚拟信标节点与监控区域未知节点交流获得的信号向量作为直推支持向量机中训练的有标签数据样本。根据训练样本的特点,提出一种多类对多类的分类法,据此推断未知节点的位置。实验与仿真结果表明,该方法获得了较高的定位精度;3.提出了一种基于节点跳数和核方法的无线传感器网络定位算法,该算法的基本思想是利用高斯核函数度量节点间的相似性。通过收集和利用实际距离和节点间跳数信息,将信标节点间的跳数信息和距离信息作为训练数据,使用偏最小二乘法学习并构建其间的最优模型,并用此模型预测未知节点到已知节点距离。实验结果表明,本定位算法定位精度较高,受信标节点的数目影响较小,并且具有将强的环境适应性,适合不同的部署环境等特点;4.提出了一种基于虚拟势场,分布式、自适应、可扩展的移动传感器网络部署算法,该算法把部署区域内的障碍物、节点等看作是带电荷的粒子,粒子受到其它障碍物和粒子的库仑力作用而产生运动,最终所有节点在力的相互作用下自动扩散到整个网络而完成部署。仿真结果表明,该算法在各种场景下的性能指标表现良好。
[Abstract]:Wireless Sensor Networks (WSNs) is produced and developed rapidly with the development of semiconductor technology, microsystem technology, communication technology and so on. Wireless sensor network is composed of all kinds of integrated micro-sensor nodes, collects and processes the data of sensing objects in the geographical area covered by the network, and processes the data through embedded system. The data is transmitted to the base station through random self-organizing wireless communication network, and finally to the management node through the Internet or satellite network. Wireless sensor networks are widely used in military defense, national security, environmental monitoring and health care. In these applications, node location and network coverage are two hot spots in wireless sensor networks. The location information of nodes plays an important role in the realization of many applications of wireless sensor networks, and network coverage determines the range of services that wireless sensor networks can provide. It also affects the cost of the network and the performance of various specific applications to a great extent. In this paper, the node location and network coverage technology in wireless sensor networks are deeply studied and discussed. In the aspect of node localization, this paper first analyzes and improves the traditional DV-Hop localization algorithm, then takes the kernel method as the research means, combines the mainstream learning algorithm, proposes two new localization algorithms, in the aspect of network coverage, In this paper, the virtual potential field based coverage algorithm in static coverage is studied, and a mobile sensor network deployment algorithm based on electrostatic field theory is proposed. The main contents and innovations of this paper are as follows: 1. In this paper, several reasons for errors in DV-Hop algorithm are extracted. On the basis of this, an incremental algorithm with coarse positioning to accurate location is proposed. The concept of collinearity is introduced into the calculation stage of trilateral measurement method, and the unit with good positioning quality is selected for calculation. Thus, the location algorithm can get a higher positioning accuracy. With the help of mobile beacon nodes, they can be moved according to the planned path, thus generating several virtual beacon nodes, which can effectively reduce the number of real beacon nodes. At the same time, the signal vectors obtained from the communication between these virtual beacon nodes and unknown nodes in the monitoring area are used as the labeled data samples trained in the direct push support vector machine. According to the characteristics of training samples, a multi-class to multi-class classification method is proposed to infer the location of unknown nodes. The experimental and simulation results show that the method has a high positioning accuracy. This paper presents a localization algorithm for wireless sensor networks based on node hops and kernels. The basic idea of the algorithm is to measure the similarity between nodes by using the Gao Si kernel function. By collecting and utilizing the information of actual distance and hops between nodes, the information of hops and distance between beacon nodes is taken as training data, and the optimal model is constructed by using partial least square method. The model is used to predict the distance between unknown nodes and known nodes. The experimental results show that the localization algorithm has the advantages of high accuracy, little influence by the number of beacon nodes, strong adaptability to the environment, and suitable for different deployment environments. This paper presents a deployment algorithm for mobile sensor networks based on virtual potential field, distributed, adaptive and extensible. The algorithm regards obstacles and nodes in the deployment area as charged particles. Particles are moved by other obstacles and particles, and eventually all nodes are automatically diffused to the entire network under the interaction of the forces. Simulation results show that the performance of the algorithm is good in various scenarios.
【学位授予单位】：南京理工大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP212.9;TN929.5

【参考文献】