摘要
无线传感网络数据流量大特征多,攻击者在无线传感网络中会利用正常数据特征隐蔽攻击行为,为了有效识别并快速阻断攻击流量,提出基于特征选择的无线传感网络攻击流量快速阻断方法。该方法使用TF-IDF算法提取流量数据特征,减少正常流量特征的干扰,使得隐蔽性特征更容易被识别,并通过信息熵增益比进行特征选择,去除特征中冗余特征向量,进一步提升网络攻击流量识别精度;再借助随机森林原理构建攻击流量识别阻断模型,精准快速识别攻击流量,并对识别出的攻击流量展开溯源阻断,保证无线传感网络的安全运行。结果表明,利用所提方法开展攻击流量阻断时,阻断准确率最高可达到99.23%,网络吞吐量为45 bit/s,数据包传输率在16 bit/s以上,具有较好的阻断效果。
Wireless sensor networks have large data traffic and many features.Attackers in wireless sensor networks will use normal data features to conceal their attack behavior.In order to effectively identify and quickly block attack traffic,a feature selection based method for quickly blocking attack traffic in wireless sensor networks is proposed.The TF-IDF algorithm is used to extract traffic data features,reduce the interference of normal traffic features,make hidden features easier to identify,and select features through information entropy gain ratio to remove redundant feature vectors,further improving the accuracy of network attack traffic recognition.By utilizing the prin-ciple of random forest,a model for identifying and blocking attack traffic is constructed to accurately and quickly identify attack traffic,and trace and block the identified attack traffic to ensure the safe operation of wireless sensor networks.The results show that the highest blocking accuracy of the proposed method can reach 99.23%,the network throughput is 45 bit/s,and the data packet transmission rate is above 16 bit/s,demonstrating good blocking effect.
作者
陈孝如
程学军
曾碧卿
CHEN Xiaoru;CHENG Xuejun;ZENG Biqing(Department of Software Engineering,Software Engineering Institute of Guangzhou,Guangzhou Guangdong 510990,China;Department of Information Engineering,Henan University of Technology Luohe Institute of Technology,Luohe Henan 462002,China;School of Software,South China Normal University,Foshan Guangdong 528225,China)
出处
《传感技术学报》
北大核心
2025年第3期526-532,共7页
Chinese Journal of Sensors and Actuators
基金
广东省普通高校特色创新科研基金项目(2021KSCX160)。
