基于一种有限队列的数据融合延时分配算法  

Delayed Allocation Algorithm Based on Limited Queue for Data Fusion

作　　者：唐懿芳[1] 钟达夫[1] 李国全[2] 赵仕俊[2] 

机构地区：[1]广东科学技术职业学院,珠海519090 [2]中国石油大学(华东),青岛266555

出　　处：《系统仿真学报》2014年第4期791-795,共5页Journal of System Simulation

基　　金：北京市教育委员会重点学科共建项目(XK100080537)

摘　　要：针对如何将网络允许延时合理地分配给数据融合簇头节点的问题,提出一种基于一种有限队列——M/G/1队列的节点延时分配算法。通过建立无线传感器网络自相似流量的数学模型,得到的网络中分组到达速率,把它作为队列的分组到达率,计算出相邻分组到达队列的间隔时间,并依此分配网络允许延时。在NS2下仿真实验表明:与级联超时法CAT和基于融合贡献的延时算法ACDA相比较,该算法有很高的融合增益,分组非实时到达率很低。With respect to the problem of how to allocate the allowable delay to fusion node reasonably, a delay distribution algorithm based on limited queue——M/G/1 queue for data fusion cluster-head nodes was proposed. The packet arrival rate which was obtained by building a mathematical model of self-similar traffic of Wireless Sensor Network（WSN） was taken as the packet arrival rate for the queue. The rate was used to calculate the adjacent packet arrival interval, and then the allowable delay can be distributed for the network. Under the simulation environment of NS2, result shows that the fusion gain is higher and the packet non-real-time arrival rate is lower than the cascade timeout method and ACDA algorithm.

关 键 词：数据融合 延时分配 自相似流量 M/G/1队列 NS2 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]