大规模无线传感器网络分布式无锚节点定位算法

摘要

This paper proposes a low-cost distributed location algorithm for large-scale wireless sensor networks — the anchor-free location algorithm which only needs to elect some reference nodes and make use of wireless transceivers carried by nodes to complete the location process. The algorithm proposed in this paper first elects five reference nodes according to certain criteria from the homogenous sensors in the network, and the five nodes are chosen as “anchors” to form a coordinate system. Other unknown nodes obtain their distances from the reference nodes by the hop gradient algorithm, which is triggered by the sink node and executed by the reference nodes one by one. The process of electing the reference nodes and the process of obtaining the distances from the reference nodes are completed simultaneously. Then the coordinate of each node is calculated by itself using Newton's iteration method. The starting value of iteration is given in this paper to further reduce the computation. This paper also gives the localization effect of the algorithm and analyzes the location error caused by node density as well.%提出了一种低成本的大规模无线传感器节点分布式定位算法--仅需要选出少量参考节点和利用节点自身的无线收发器即可实现整个网络节点的定位的无锚节点定位算法.该算法首先从网络中所有普通节点中按照一定准则选取5个基准节点,以这5个基准节点作为"锚节点",确定一个坐标系;其他未知节点将根据到基准节点的距离计算出自身坐标.未知节点到基准节点之间的距离通过基准节点建立的跳数梯度场求得.跳数梯度场的建立过程由核心节点触发开始,其他基准节点依次执行.基准节点的选取过程与得到各未知节点到基准节点的距离的过程同时进行.未知节点根据自己到各基准节点的距离,利用牛顿迭代法计算出自身的坐标.为了进一步降低计算量,给出了迭代初值的选取方法.通过仿真实验给出了利用该算法得到的定位效果,并对误差影响进行了对比分析.