移动无线传感器网络中基于矩阵的定位算法研究

发布时间：2019-04-18 13:40

【摘要】：无线传感器网络(Wireless Sensor Networks,WSNs)是指由大量的静止或移动的传感器节点以自组织和多跳的方式构建的无线网络,已经应用于众多场合,如环境监测、目标跟踪、交通控制、人体健康监测、军事以及救灾等领域。但是这些应用都需要传感器节点知道自身的位置信息,脱离位置信息的感知数据对于大多数应用是没有价值的。而通过人工部署传感器网络中的节点或为需要位置信息的节点安装GPS设备都会受到传感器节点自身特点或应用场景的限制。因此,需要传感器节点通过运行定位算法或其他机制得到自己的位置信息。现有的无线传感器网络节点定位算法普遍存在着定位误差较大、算法复杂度高、通信消耗大、不适用于移动节点等其中一项或多项问题。随着无线传感器网络技术的不断成熟和应用的不断广泛,针对移动传感器网络的实用化节点定位技术具有重要的理论意义和应用价值。本文针对移动传感器网络中的节点定位问题主要完成了如下工作:本文首先分析概述了研究问题的背景和意义,并总结了现有的定位测距技术、定位原理以及现有定位算法的优缺点。其次,在此基础上,本文提出了一种基于矩阵填充的分布式定位算法MALL(Matrix-completion Localization)。MALL算法是利用一系列的约束条件,如节点间距离约束、节点坐标具有低秩性、时间稳定性等,来求解目标函数最优化的分布式定位算法,保证了算法较高的定位精度和易扩展性的优势。由于MALL算法只涉及到凸优化和低复杂度的非凸优化计算,算法计算复杂度较低,能够实现移动网络节点的快速定位。MALL算法仅使用一跳普通邻居节点信息和两跳锚定邻居信息来完成定位过程,具有较低的定位通信消耗。在本文中,对MALL算法的计算复杂度进行了理论分析。最后,通过仿真实验对MALL算法与其他一些现有算法进行了定位精度、算法运行时间、通信消耗性能方面的比较。实验表明,MALL算法优于现有算法。
[Abstract]:Wireless sensor network (Wireless Sensor Networks,WSNs) is a wireless network constructed by a large number of static or mobile sensor nodes in a self-organized and multi-hop manner. It has been used in many situations, such as environmental monitoring, target tracking, traffic control, and so on. Human health monitoring, military and disaster relief. However, these applications require sensor nodes to know their own location information, and sensing data without location information is of no value to most applications. However, the manual deployment of nodes in sensor networks or the installation of GPS devices for nodes that require location information will be limited by the sensor nodes' own characteristics or application scenarios. Therefore, sensor nodes need to run location algorithm or other mechanisms to obtain their own location information. The existing localization algorithms for wireless sensor networks generally have large positioning errors, high complexity and large communication consumption, so they are not suitable for one or more of the problems such as mobile nodes. With the development and wide application of wireless sensor networks (WSNs), the practical node positioning technology for mobile sensor networks (MSNs) is of great theoretical significance and application value. The main work of this paper is as follows: firstly, the background and significance of the research are analyzed and summarized, and the existing location and ranging techniques are summarized. Location principle and advantages and disadvantages of existing location algorithms. Secondly, a matrix-filled distributed localization algorithm (MALL (Matrix-completion Localization) is proposed in this paper. Mall uses a series of constraints, such as distance constraints between nodes, and the coordinates of nodes are of low rank. Time stability and so on, to solve the objective function optimization of the distributed location algorithm, to ensure the algorithm high positioning accuracy and easy to expand the advantages of the algorithm. Because the MALL algorithm only involves convex optimization and non-convex optimization with low complexity, the computational complexity of the algorithm is low. Mall algorithm only uses one-hop common neighbor node information and two-hop anchor neighbor information to complete the positioning process, which has a low location communication consumption. In this paper, the computational complexity of MALL algorithm is analyzed theoretically. Finally, the positioning accuracy, running time and communication consumption performance of the MALL algorithm are compared with other existing algorithms through simulation experiments. Experimental results show that the MALL algorithm is superior to the existing algorithms.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP212.9;TN929.5

【参考文献】