基于竞争学习的无线传感器网络聚类算法

在复杂的煤矿作业环境下，无线传感器网络面临着从节点组织、灾害发生后网络重组、救援人员所携带的传感器节点加入网络等问题。该文提出一种基于竞争学习的无线传感器网络聚类算法，归一化处理传感器节点的剩余能量和邻居节点个数，通过相似性比较实现节点间竞争。仿真表明，该算法通信能耗低，网络能耗均衡，扩展性好，抗毁能力较强。

Clustering Algorithm Based on Competitive Learning in Wireless Sensor Networks

In the complex environment of coal mine, there are some challenges in wireless sensor networks such as organization of following nodes, rebuilding of the network after disaster and adding of nodes taken by rescuers. A clustering algorithm is proposed based on competitive learning, in which the remaining energy and the number of neighboring nodes for sensor node are normalized and the competition between nodes are achieved by similarity comparison. Simulation results show that the algorithm represents low communication energy consumption, balance of network energy consumption, better expansibility and stronger ability to resist destruction.