无线传感器网络中的分布式节点定位方法

提出一种基于流形学习的分布式Hessian局部线性嵌入（DHLLE）定位方法,给出了基于流形学习算法的定位框架.DHLLE方法采用同情最邻近算法来选择节点邻居列表,并应用Hessian局部线性嵌入（HLLE）算法获取传感器网络节点的局部映射,再通过对局部映射合并获得所有节点的全局映射,最后通过对参考节点进行坐标匹配以取得所有节点的全局坐标.仿真结果表明,DHLLE方法能够快速、准确地对节点进行定位,且复杂度低,节点能耗小,其性能超过了分布式加权多维定标等算法.

Distributed Nodes Localization Method in Wireless Sensor Networks

A novel distributed Hessian local linear embedded(DHLLE) localization method and framework based on manifold learning were presented,in which the algorithm of sympathy the nearest was used to select the neighbors list,and the HLLE algorithm was used to obtain the local map of sensor network's nodes.Then,by combining local maps the global map of all nodes was acquired.Finally,the global coordinates of all nodes could be obtained through matching to reference coordinates.The simulation results demonstrate that DHLLE can localize the nodes accurately and rapidly with lower complexity and less energy consumption in nodes,and its performance is superior to the distributed weighted multi-dimension localization algorithm.