基于灰色预测的分布式传感器网络故障检测方法  

A Gray-Forecast Algorithm of Fault Detection for Wireless Sensor Networks

作　　者：张健[1,2] 

机构地区：[1]北京工业大学嵌入式软件与系统研究所,北京100124 [2]北京建筑大学理学院,北京100044

出　　处：《传感技术学报》2015年第8期1188-1193,共6页Chinese Journal of Sensors and Actuators

基　　金：北京建筑大学科学研究基金项目(00331614047)

摘　　要：针对无线传感器网络节点故障原因复杂,基于灰色预测理论,提出一种故障检测方法。算法通过分析在某一采样间隔内,观测节点数据变化趋势是否与邻居节点变化趋势一致,从而确定节点是否异常。仿真实验表明,算法故障检测能力较强,且避免了节点瞬间失效情况的出现。同时该算法设计简单,易于硬件实现。Wireless Sensor Networks （WSNs） are widely used in environmental monitoring,health care,smart home, military fields,and so on,because of its small,flexibility,low cost and unattended. And faults are common due to the bad environment and unattended. In order to ensure WSNs’service is normal,it is necessary to detect the faults. According to the complex reasons of wireless sensor network nodes failure,this paper presents a new fault detection algorithm based on based on gray prediction theory. The algorithm is that within a sample interval,it is analyzed that the observed node’s data trend is consistent with the trend of changed in the neighbour nodes to determine whether the node is abnormal. The experiment shows fault detection precision is very high,and the algorithm avoid the node instant failure. Meanwhile,the algorithm is simple,easily implemented in hardware.

关 键 词：无线传感器网络 故障检测 灰色预测 分布式 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]