一种面向无线传感器网络相对定位的分簇算法

Clustering algorithm for relative localization in wireless sensor network

测距误差累积是影响无线传感器网络相对定位算法性能的主要因素之一,网络分簇是降低这一误差的有效手段。针对相对定位特点,基于典型分簇算法——OK(Overlapped K-hop),提出EOK(Enhanced Overlapped K-hop)分簇算法。EOK算法改进了OK算法的簇头节点选择机制,提出邻居簇头节点合并机制,使得节点分簇更加符合定位应用需要。仿真实验表明,相比OK算法,采用EOK算法产生的节点簇数量更少、节点簇分布更均匀,在多数网络条件下具有更低的算法通信开销。

Ranging accumulation error is one of the main factors affecting the performance of relative localization in wireless sensor network,and can be reduced through clustering effectively.A new clustering algorithm-EOK （Enhanced Overlapped K-hop） which is based on OK （Overlapped K-hop） clustering algorithm is proposed in this paper.It enhances the choosing mechanism of cluster nodes in OK,and presents a combination mechanism of neighbor cluster nodes.Clustering is more suitable for relative localization using EOK.In comparison with OK,simulation results show that EOK provides less number of cluster nodes and cluster＇s distribution is well-proportioned.Moreover,it can reach a lower level of the communication cost in many network conditions.