Abstract: Due to lack of network stability considerations, most existing Ad Hoc clustering algorithms are difficult to apply in the actual scene. At first, analysis and comparison of three algorithms form existing clustering algorithms, then on the basis of fuzzy clustering, the paper designs Packet data packets and timers, and also uses the concept of membership degree, at final proposes a new clustering algorithm. Simulation results show that compared with LOW-ID algorithm, WCA algorithm and LEACH algorithm, this algorithm has better stability.

Key words: Ad Hoc network, clustering algorithm, fuzzy clustering, membership μ

摘要： 现有的Ad Hoc网络分群算法大多对网络的稳定性考虑不足，在实际场景中难以应用。从已有分群技术出发，对三种分群算法进行比较分析，在模糊聚类的基础上，设计了Packet数据包和定时器，并利用隶属度的概念，提出了一种新型CABF（Clustering Algorithm Based FCM）分群算法。仿真结果表明，该算法与最小ID算法、WCA算法、LEACH算法相比，具有更好的稳定性。

关键词: Ad Hoc网络, 分群算法, 模糊聚类, 隶属度&mu