分布式信息融合的物联网事件检测方法  

Distributed Event Detection for the Internet of Things

作　　者：楼晓俊[1] 鲍必赛[1] 刘海涛[1,2] 

机构地区：[1]中国科学院上海微系统与信息技术研究所无线传感器网络与通信重点实验室,上海200050 [2]无锡物联网产业研究院,江苏无锡214135

出　　处：《南京邮电大学学报（自然科学版）》2012年第1期12-16,共5页Journal of Nanjing University of Posts and Telecommunications：Natural Science Edition

基　　金：国家重点基础研究发展计划(973计划)(2011CB302906);国家科技重大专项(2010ZX03006-004)资助项目

摘　　要：事件检测是物联网感知技术的一项重要应用。由于感知节点资源受限,事件检测准确率与节点能量有效性较难平衡。提出了一种新的物联网事件检测方法,基于分布式无线传感器网络,采用特征选择与人工智能学习算法,对感知数据进行局部与全局的二级信息融合,在降低节点功耗的同时提升了检测的准确率。进行要地周界入侵事件检测实验,结果表明,该方法相比传统的事件检测方法更加有效。In sensing technology of the Internet of Things （IoT） ,event detection is one of the most impor- tant applications. However,there is always a hard tradeoff between the detection accuracy and the power efficiency due to limited resources of the sensor nodes. A novel event detection method for IoT is proposed based on the distributed wireless sensor network. A twostep data fusion model of applying feature selection and artificial intelligence algorithms is built for event detection to both reduce power consumption and improve detection accuracy. Experiment results of perimeter intrusion detection show that the proposed method can be more effective than traditional event detection methods.

关 键 词：物联网 事件检测 分布式传感网 特征选择 数据融合 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]