基于改进FastICA算法的工控网络恶意节点攻击检测系统设计  

Design of Malicious Node Attack Detection System in Industrial Control Network Based on Improved FastICA Algorithm

作　　者：江山杉 JIANG Shanshan(Hospital of Chengdu University of Traditional Chinese Medicine,Chengdu 610032,China)

机构地区：[1]成都中医药大学附属医院,四川成都610032

出　　处：《微型电脑应用》2024年第9期49-53,共5页Microcomputer Applications

基　　金：四川省自然科学基金(2020SC09341)。

摘　　要：为了确保工控系统的安全和可靠性,基于改进FastICA算法,设计工控网络恶意节点攻击检测系统。在工控网络服务器端集群中,从网络服务器端、攻击行为排重模块、恶意节点辨识模块和Fabric攻击行为检测组件等4部分设计工控网络恶意节点攻击检测系统的硬件模块。在此基础上计算Fast初始值,统计ICA单位的不确定性,改进FastICA算法。由此,联合相关攻击行为信息,推导恶意节点的攻击分离条件。实验结果表明,按照1∶1的负荷标准分配待测数据,所设计系统可以有效检测恶意节点攻击行为,得到的恶意节点实时检测数量值与额定值的拟合度接近100%,能够较好维护网络体系的运行稳定性。To ensure the safety and reliability of the industrial control system,this paper designs a malicious node attack detection system in industrial control network based on improved FastICA algorithm.In the industrial control network server cluster,the hardware module of malicious node attack detection system in industrial control network is designed from four parts,network server,attack behavior de-duplication module,malicious node identification module and Fabric attack behavior detection component.On this basis,this paper calculates the initial value of Fast,tallies the uncertainty of ICA units,and improves the FastICA algorithm.Therefore,combining relevant attack behavior information,the attack separation conditions for malicious node are derived.The experimental results show that by distributing the test data according to a 1∶1 load standard,the designed system can effectively detect malicious node attack behavior,and the obtained real-time detection quantity value of malicious node has a fitting degree close to 100%with the rated value,which can better maintain the operational stability of the network system.

关 键 词：改进FastICA算法 工控网络 恶意节点 攻击检测 节点辨识 节点分离 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]