云计算环境下船舶监控网络异常数据检测方法  

Abnormal data detection method of ship monitoring network in cloud computing environment

作　　者：农嘉[1] 王代远[1] 潘梅勇[1] 覃志松[2] NONG Jia;WANG Dai-yuan;PAN Mei-yong;QIN Zhi-Song(Guangxi Eco-engineering Vocational and Technical College,Liuzhou 545000,China;Guilin University of Electronic Technology,Guilin 541004,China)

机构地区：[1]广西生态工程职业技术学院,广西柳州545000 [2]桂林电子科技大学,广西桂林541004

出　　处：《舰船科学技术》2021年第8期190-192,共3页Ship Science and Technology

基　　金：广西职业教育教学改革研究重点项目(GXGZJG2018A017);广西生态职院校级项目(2019KY10)

摘　　要：在船舶监控网络高度应用的今天,监控网络异常数据检测受到了学术界以及船舶制造业的高度关注。就目前的船舶监控网络异常数据检测方法而言,其计算能力较低,导致异常数据误报率较高。针对此问题,设计云计算环境下船舶监控网络异常数据检测方法。使用相似度函数对监控节点数据展开相似性检测,初步确定数据异常节点位置。根据节点位置,对监控网络数据进行时间序列检测,确定异常数据输出时间。对上述两部分进行融合处理,完成异常数据检测方法的设计过程。经对比实验验证可知,此方法在应用中具有误报率低,计算效率较高的优点,可将其应用到后续的船舶监控网络数据处理过程中。Today,with the highly application of ship monitoring network,abnormal data detection of monitoring network has been highly concerned by academia and shipbuilding industry.As far as the current abnormal data detection methods in ship monitoring network are concerned,their computing power is low,which leads to high false alarm rate of abnormal data.In order to solve this problem,this paper designs the abnormal data detection method of ship monitoring network in cloud computing environment.The similarity function is used to detect the similarity of monitoring node data,and the location of abnormal data node is preliminarily determined.According to the node location,the time series of monitoring network data is detected to determine the abnormal data output time.The above two parts are fused to complete the design process of abnormal data detection method.The experimental results show that this method has the advantages of low false alarm rate and high computational efficiency,which can be applied to the subsequent ship monitoring network data processing.

关 键 词：异常数据检测 监控网络 云计算 相似度计算 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]