强噪声干扰下大规模网络入侵信号提纯仿真  

Large-scale Network Intrusion Signal Purification Simulation under Strong Noise Interference

作　　者：匡凤飞[1] KUANG Feng-fei(Minnan Science and Technology Institute,Fujian Normal University,Quanzhou Fujian 362332,China)

机构地区：[1]福建师范大学闽南科技学院,福建泉州362332

出　　处：《计算机仿真》2018年第9期291-294,共4页Computer Simulation

基　　金：2017年教育部-华为技术有限公司产学合作协同育人项目(201701001020)

摘　　要：针对目前强干扰噪声下网络入侵信号提纯方法精度低、漏检率高的现象,提出基于检测阈值的强噪声干扰下大规模网络入侵信号提纯方法。将自适应噪声抵消体系划分主通道、参考通道、两个传感器,主通道接收从一路传感器传来的有用信号与干扰噪声信号形成的混合信号。采用参考通道对另一路传感器传输来的噪声进行检测。基于滤波器利用减法器实现主通道噪声分量最大程度抵消,输出可利用信号。将该信号代入网络入侵信号提纯中,依据经验模态分解法计算网络入侵剩余干扰信号的时域,得到剩余噪声信号通过网络数据接收设备混频之后的离散值。设定剩余噪声信号能量分布可用熵,以此获取定频信号检测函数,确定检测门限阈值,根据该阈值得到强噪声干扰下网络入侵信号提纯结果。实验结果表明,上述方法下入侵信号提取系数最高为0. 96,漏检率为1. 25%。所提方法具有较高的运行精度和查全性能,可实践性较强。At present, the network intrusion signal purification method in strong noise has low accuracy and high missed rate. Therefore, a method to purify large-scale network intrusion signal in strong noise interference based on detection threshold value was proposed. Firstly, the self-adaptive noise cancellation system was divided into a main channel, a reference channel and two sensors. Secondly, the main channel received the mixed signal formed by the useful signal sent from one sensor and the interference noise signal. Then, reference channel was used to detect the noise transmitted from another sensor. Based on the filter, the subtracter was used to achieve the maximum cancellation of noise component in main channel and output available signal. Moreover, this signal was introduced into the network intrusion signal purification and the empirical mode decomposition method was used to calculate time domain of residual network intrusion interference signals. Then, the discrete value of remaining noise signal after the mixing of network data receiving equipment was obtained. Finally, the available entropy of energy distribution of residual noise signals was set to obtain the detection function of fixed-frequency signal and determine the detected threshold value. Thus, the result of network intrusion signal purification in strong noise interference was obtained based on threshold value. Simulation results show that the proposed method has high running accuracy and recall performance, and the maximum coefficient of intrusion signal extraction is 0.96 and the missed detection rate is 1.25%.

关 键 词：强噪声干扰 网络入侵信号 提纯 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]