移动无线传感器网络中分布式重聚类算法研究

移动无线传感器网络中,节点的移动性影响着层次化聚类之后的网络结构,从而影响聚类内部节点间通信时的数据送达率与能耗.为了降低节点移动性的影响,本文提出了一种分布式重聚类算法.该算法基于已聚类网络,利用粒子滤波算法对节点当前位置进行估计,并结合移动模型预测下一时刻位置;处于聚类边界的非簇头节点周期性地评估自身是否需要重聚类,并在需要时通过与所属聚类及目标聚类的簇头节点通信,将自身重聚类到目标聚类中.仿真结果表明,在重聚类周期较小时,该算法能够使节点在移动过程中保持合理的通信距离,并在数据送达率与能耗方面优于现有的算法.

Study on distributed re-clustering algorithm for moblie wireless sensor networks

In mobile wireless sensor networks,node mobility influences the topology of the hierarchically clustered network,thus affects packet delivery ratio and energy consumption of communications in clusters.To reduce the influence of node mobility,a distributed re-clustering algorithm is proposed in this paper.In this algorithm,basing on the clustered network,nodes estimate their current locations with particle algorithm and predict the most possible locations of next time basing on the mobility model.Each boundary node of a cluster periodically estimates the need for re-clustering and re-cluster itself to the optimal cluster through communicating with the cluster headers when needed.The simulation results indicate that,with small re-clustering periods,the proposed algorithm can be effective to keep appropriate communication distance and outperforms existing schemes on packet delivery ratio and energy consumption.