基于无线传感器网络的K均值算法研究

　　传统无线传感网一般由大量密集的传感器节点构成，存在节点计算能力、能源和带宽都非常有限的缺点，为了有效节能、延长网络寿命，介绍了基于聚类的K均值算法。该算法通过生成的簇头节点散播到网络的各个区域中，减少了每个区域内通信的能耗和可能会出现的一般节点过早死亡的情况，从而避免了网络对该区域提早失去监控。实验证明，该算法对各节点位置确定的无线传感器网络有低能耗、高稳定性特点，能较好的对无线传感器网络节点拓扑结构进行优化，达到能量均衡的目的。

　　Abstract:

　　Traditional wireless sensor network was consisted of a large number of dense sensor nodes generally， which has the disadvantage that the node’s computing power， energy and bandwidth are very limited. In order to save energy effectively and prolong the network lifetime， the K-means algorithm based on clustering was introduced. By interspersing generated cluster head nodes to each area of network， the algorithm reduced the communication consumption of each area and the possible condition of common nodes’ premature death. Thus it avoid the network prematurely losing control to the region. The experiment results show that for the wireless sensor networks that each node has determinate location， the algorithm has the features of low energy consumption and high stability. It can preferably optimize the topology of wireless sensor network nodes， and achieve the purpose of energy equilibrium.