基于Fiedler矢量的分布式自适应分簇算法

Distributed adaptive clustering algorithm based on Fiedler vector

针对无线传感器网络分簇(clustering)问题,提出一种基于Fiedler矢量的分布式分簇改进算法。该算法利用Fiedler矢量的元素符号特性对网络进行递归分簇处理,引入网络拓扑信息,根据网络自身的内部连接自适应决定分簇数目,通过Fiedler矢量的元素数值选出簇头,并且算法给簇头子集筛选合适的网关节点以确保簇头子集的连通性。仿真实验表明,在共识频谱感知的基础上,该算法生成的簇头子集与全网络共识所收敛的结果相同,簇头子集共识收敛速度相对更快,耗时短,能够以更好的时效性、更高的能效达到与全网络共识收敛相同的效果。

To solve the problems of node clustering in the wireless sensor network, an improved distributed adaptive cluste-ring algorithm based on Fiedler vector is proposed. The algorithm utilizes the positive and negative characteristics of the Fiedler vector element to cluster recursively, and leads into the network topology information, then determining the number of clustering adaptively in line with its own internal network connections. The algorithm filters out the cluster head through the Fiedler vector element values. In addition, it adds appropriate gateway nodes for head nodes to ensure that the cluster heads set is connected. Simulation results show that on the basis of the consensus cooperative spectrum sensing, there, s lit-tle difference between the result of the cluster heads set consensus and that of the whole network consensus, but the cluster heads set consensus converges faster and gets a shorter time-consuming. It can achieve the same performance as the the whole network algorithm with better real-time performance and higher energy consumption.