摘要
无线传感网络因其自身的拓扑结构具有不确定性,易受到恶意攻击,常规方法没有考虑的链路特征,导致网络的安全性较低。为此,提出一种基于链路特征的无线传感网络数据多步攻击识别方法。通过支持向量分类机对无线传感网络展开链路安全评估,得到链路安全等级,提取链路特征,获取传感器网络内部链路特征分布情况。引入自组织映射神经网络提取无线传感网络数据多步攻击特征,建立无线传感网络数据多步攻击特征组合,结合链路特征分布情况,通过特征组合识别无线传感网络数据多步攻击。仿真结果表明,采用所提方法的无线传感网络数据多步攻击识别率高于96%,误报率低于0.22%,识别时间在24 ms以内,具有良好的识别效果。
Wireless sensor networks are vulnerable to malicious attacks due to their uncertain topology.The link characteristics that are not considered in conventional methods lead to low network security.Therefore,a multi step attack identification method based on link characteristics for data in wireless sensor networks is proposed.The Support Vector Machines is used to evaluate the link security of wireless sensor networks.The link security level is obtained,the link characteristics are extracted,and the distribution of link characteristics within the sensor network is obtained.Self-organizing mapping neural network is introduced to extract multi-step attack characteristics of wireless sensor network data,a multi-step attack feature combination of wireless sensor network data is established,and the distribution of link characteristics is combined to identify multi-step attacks of wireless sensor network data through the feature combination.The simulation results show that the proposed method has a good recognition effect with a multi-step attack identification rate of more than 96%,a false alarm rate of less than 0.22%,and a recognition time of less than 24 ms.
作者
徐欢
李洁珊
零颖俏
杨春
XU Huan;LI Jieshan;LING Yingqiao;YANG Chun(School of Computer Science and Technology,Huazhong University of Science and Technology,Wuhan Hebei 430000,China;China Southern Power Grid Co.,Ltd.,Guangzhou Guangdong 510000,China;China Southern Power Grid Energy Development Research Institute,Guangzhou Guangdong 510000,China)
出处
《传感技术学报》
CAS
CSCD
北大核心
2024年第11期1958-1963,共6页
Chinese Journal of Sensors and Actuators
基金
国家重点研发计划项目(2021YEE0204200)
南方电网公司信息专题研究项目“数字电网理论研究”(000000HY41220007)
南方电网公司信息专题研究项目“电力数据要素合作生态体系构建与培育研究”(000000HY41220003)。
关键词
无线传感网络数据
多步攻击识别
链路特征
支持向量分类机
自组织映射神经网络
wireless sensor network data
multistep attack identification
link characteristics
support vector machines
self organizing mapping neural network
作者简介
徐欢(1984-),男,汉族,广东梅州人,硕士,高级工程师,毕业于华中科技大学计算机科学与技术学院,主要从事大数据分析、数据资产管理、数据安全等研究工作,现就职于中国南方电网有限责任公司,担任高级工程师,发表过多篇论文,xg6587444@yeah.net;李洁珊(1985-),女,汉族,硕士,广东潮州人,高级工程师,研究方向:大数据、数字孪生;零颖俏(1994-),女,汉族,广西防城港人,工程师,研究方向:大数据分析;杨春(1984-),男,汉族,湖北宜都人,博士,高级工程师,研究方向:大数据、人工智能。
