无线传感器网络三维表面定位算法研究

发布时间：2018-06-23 10:35

本文选题：节点定位 + 凹凸节点　；参考：《电子科技大学》2014年硕士论文

【摘要】：随着近年来无线传感器网络在军事及民用的应用越来越广泛,人们越来越关注无线传感器网络在各方面的研究。其中节点定位技术是传感器网络应用的关键支撑技术之一,针对节点定位技术的研究面也越来越广,从二维平面定位到三维表面定位,从固定锚节点到移动锚节点的定位研究。其中,三维表面定位最接近无线传感器网络的实际应用,然而国内外对其的研究相对比较少,并且缺乏考虑到实际的复杂地形如凹凸表面对其定位误差和节点定位率的影响。本文在二维定位算法的基础上提出了三维表面定位的复杂地形对节点定位的影响,主要有凹凸地形对节点定位精度和节点定位率的影响,以及锚节点的分布不均匀对节点定位和节点能耗的影响。针对这些问题,本文首先提出了三维表面凹凸分解的分层定位算法,该算法首先将网络基于高度分层,从而减小在应用场景比较大时减小节点定位误差的迭代,然后在每层内根据凹凸节点划分子网,将凹凸度相近并相邻的节点划分到一个子网内,在每个子网内定位出节点的坐标,最后合并。该算法在定位误差和定位率方面都比其余的算法有所改进,不过计算复杂度和能耗方面提高很小。所以本文又提出了基于关键节点的移动路径规划定位算法,该算法是在三维表面凹凸分解定位算法的基础上进行了改进,主要包含以下几方面:(1)该算法将网络中的节点映射到一个平面上,通过分辨是否有部分区域重合来判断网络是否需要分割,如果有重合将网络划分成多个子网,从而提高节点的定位率。(2)该算法根据凹凸节点的定义在子网内找出所有凹凸节点定义为关键节点,并使用移动锚节点定位出关键节点并作为已知节点,然后利用这些已知节点定位其余的未知节点,避免凹凸节点对定位误差的影响。(3)算法采用移动锚节点替换固定锚节点,可以适应各种复杂的地形,避免了锚节点密度不够或者分布不均匀导致的节点不可被定位和定位误差迭代较大的情况。本文在最后通过实验证明两个算法在定位误差,定位率和节点能耗方面都有较大的改进,并且后者比前者在各方面都有明显的提升。
[Abstract]:With the application of wireless sensor networks in military and civilian applications, more and more attention has been paid to the research of wireless sensor networks. The node location technology is one of the key supporting technologies in sensor network application. The research area of node location technology is more and more extensive, from two-dimensional plane positioning to three-dimensional surface positioning, from fixed anchor node to mobile anchor node location research. Among them, 3D surface localization is most close to the practical application of wireless sensor networks, but the research on it is relatively few at home and abroad, and lack of consideration of the actual complex terrain such as concave and convex surface on its positioning error and node localization rate. In this paper, based on the two-dimensional localization algorithm, the influence of the complex terrain of 3D surface location on node location is proposed, including the effect of concave and convex terrain on node location accuracy and node localization rate. And the influence of uneven distribution of anchor nodes on node location and node energy consumption. To solve these problems, this paper first proposes a layered localization algorithm based on three dimensional surface concavity and convex decomposition. Firstly, the network is based on high stratification, which reduces the iteration of node localization error when the application scene is large. Then the subnets are divided into subnets according to the concave and convex nodes in each layer, and the adjacent nodes with similar concave and convex degrees are divided into a subnet, and the coordinates of the nodes are located in each subnet, and finally the nodes are merged. The algorithm improves the localization error and localization rate compared with the other algorithms, but the computational complexity and energy consumption are little improved. Therefore, this paper proposes a mobile path planning and location algorithm based on key nodes, which is improved on the basis of 3D surface concavity and convex decomposition localization algorithm. It mainly includes the following aspects: (1) the algorithm maps the nodes in the network to a plane, and determines whether the network needs to be divided by distinguishing whether there is a partial area overlap. If there is overlap, the network is divided into several subnets. In order to improve the localization rate of nodes. (2) according to the definition of concave and convex nodes, the algorithm finds out that all concave and convex nodes are defined as key nodes in the subnet, and uses mobile anchor nodes to locate the key nodes and take them as known nodes. Then using these known nodes to locate the remaining unknown nodes to avoid the impact of concave and convex nodes on the positioning error. (3) the algorithm uses mobile anchor nodes to replace fixed anchor nodes, which can adapt to various complex terrain. It avoids the problem that the node can not be located and the error iteration is large due to the lack of anchor node density or the uneven distribution. At the end of this paper, it is proved by experiments that the two algorithms have great improvement in location error, localization rate and node energy consumption, and the latter has obvious improvement over the former in all aspects.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.5;TP212.9

【参考文献】