摘要
针对传统攻击信号监测方法监测时数据迭代次数过多,占用资源的问题,研究基于类神经网络的嵌入式系统攻击信号监测方法。利用信号的相关性特性,检测攻击嵌入式系统的信号。在时域中提取检测到的信号的频率特征,并在循环谱中提取攻击信号的循环谱特征。根据提取的信号特征建立特征库,设计类神经网络分类器。使用训练集训练类神经网络分类器,确定分类器参数,实现对攻击信号的监测。通过与传统攻击信号监测方法的对比实验,证明了在完成相同工作任务量的情况下,基于类神经网络的嵌入式系统攻击信号监测方法的数据迭代次数更少,占用资源更少,性能更佳。
In view of the problem that the traditional attack signal monitoring method has too many data iterations and takes up resources,the attack signal monitoring method of embedded system is studied in this paper based on neural network.The signal of attacking embedded system is detected by using the correlation characteristic of the signal.The frequency characteristics of detected signals are extracted in the time domain and the cyclic spectrum characteristics of attack signals are extracted in the cyclic spectrum.Based on the extracted signal features,a feature library is established and a neural network classifier is designed.The training set is used to train the neural network classifier,and the classifier parameters are determined to monitor the attack signal.By comparing with the traditional attack signal monitoring method,it is proved that the embedded system attack signal monitoring method based on neural network has fewer data iterations,less resources and better performance when the same amount of work is completed.
作者
黄煜坤
陈珩
HUANG Yukun;CHEN Heng(Guangxi Vocatioruil&Technical Institute of Industry,Nanning Guangxi 530001,China)
出处
《自动化与仪器仪表》
2020年第7期50-53,共4页
Automation & Instrumentation
基金
2016年度广西高校中青年教师基础能力提升项目(No.2016003YB009)资助。
关键词
类神经网络
嵌入式系统
攻击信号监测
信号特征提取
neural networks
embedded system
attack signal monitoring
signal feature extraction
