摘要
考虑到无线传感器网络中传感器节点经常部署在恶劣的环境中以及节点自身资源具有有限性,节点在运行的过程中容易产生错误数据,造成漏警和虚警错误,从而影响网络的正常运行.本文提出了一种基于预测机制的事件检测容错算法.首先利用事件的时空关联性对监测区域进行事件检测,对事件是否发生还是节点异常进行准确判断.然后针对错误的数据进行具体分析,当节点出现异常的时候,利用KNN-PSOELM预测机制进行估计,以排除错误数据对融合结果的影响,执行有效的容错算法.因此,准确地区分节点数据的类型,排除错误数据的影响,提高事件检测的准确度.
The harsh environment and the limited resources of nodes in wireless sensor network are prone to error data which can result in missed alarm and false alarm errors in the process of running.The missed alarm and false alarm errors have a bad effect on the quality of the network.Therefore,a prediction-based fault-tolerant aggregation algorithm was proposed.Firstly,the events were detected by using temporal-spatial correlation between events and determine whether the event happened or the node was abnormal.When the node is out of order,a specific analysis of faulty data that may occur during data gathering was made and different fault-tolerant solutions were pointedly put forward using the method （K-Nearest Neighbor-Particle Swarm Optimization-Extreme Learning Machine,KNN- PSOELM） based prediction model.Thus,we can effectively improve the event detection accuracy,distinguish different data types and eliminate the influence of faulty data.
作者
刘耿耿
郭文忠
洪伟
LIU Geng-geng1,2 GUO Wen-zhong1,2 ,HONG Wei1(1 College of Mathematics and Computer Sciences, Fuzhou University, Fuzhou 350116, China; 2 Key Laboratory of Network Computing and Intelligent Information Processing （Fujian Province }, Fuzhou University, Fuzhou 350116, China)
出处
《小型微型计算机系统》
CSCD
北大核心
2018年第4期787-792,共6页
Journal of Chinese Computer Systems
基金
国家重点基础研究发展计划项目(2011CB808000)资助
国家自然科学基金项目(11501114
11271002
61672159)资助
关键词
无线传感器网络
预测机制
事件检测
容错
时空关联性
wireless sensor networks
prediction mechanism
event detection
fault-tolerant
temporal-spatial correlation
