基于信誉评估的Ad hoc分布式入侵检测  

Distributed Intrusion Detection of Ad hoc Network Based on Reputation

作　　者：陈赵懿 高秀峰 王帅 韩月明 CHEN Zhao-yi;GAO Xiu-feng;WANG Shuai;HAN Yue-ming(Equipment Simulation Training Center,Shijiazhuang Campus of Army Engineering University,Shijiazhuang 050003,China)

机构地区：[1]陆军工程大学石家庄校区模拟训练中心,石家庄050003

出　　处：《火力与指挥控制》2020年第9期6-11,共6页Fire Control & Command Control

基　　金：国家自然科学基金(61271152);军内科研基金资助项目(130105197302074514)。

摘　　要：为提高节点资源受限条件下的Ad hoc网络入侵检测率,提出了基于信誉评估的分布式入侵检测模型,模型分为信誉评估、贝叶斯博弈检测和综合检测三部分。采用邻居节点共享信誉的方式,综合计算本地直接信誉与邻居节点的间接信誉,提高信誉评估准确性。根据信誉评估值选择检测方式,贝叶斯博弈检测动态调整检测策略可节省资源,综合检测可提高入侵检测率和降低误检率。通过实验与数据分析,本模型具有较高检测率和较低的资源消耗。In order to improve the intrusion detection rate under the condition of limited resources of Ad hoc network nodes,a distributed intrusion detection model based on reputation is proposed in this paper.The model is divided into three parts,reputation evaluation,Bayesian game detection and compound detection.The direct reputation of the local node and the indirect reputation of the neighbor nodes are calculated comprehensively,thus improving the accuracy of the reputation evaluation.Then,the detection method is selected based on the reputation value of the detected node.When the reputation of the detected node is high,Bayesian game detection is choosen to reduce the consumption of node resources.On the contrary,when the node reputation decreases and exceeds the set threshold,compound detection to is selected improve detection rate.Through data analysis,this model has high detection rate and low resource consumption.

关 键 词：AD HOC 信誉 贝叶斯博弈 入侵检测 交叉验证 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]