摘要
针对无线传感器网络中的攻击,提出一种无线传感器网络的两级入侵检测模型.采用主成分分析法进行特征降维,降低数据存储量和计算量;簇级中普通节点采用K近邻直推式信度机进行异常检测,簇头采用粒子群优化参数的支持向量机对检测到的异常进行进一步误用检测分类,保障节点安全;基站级将异常检测技术与误用检测技术相结合,处理簇头提交的监测数据,可同时提高检出率和降低误报率,保障簇头安全.仿真结果显示本文算法在小样本情况下能够提高检测正确率.
A two-level intrusion detection model of wireless sensor networks （WSNs） is proposed to detect the attacks in WSNs. The principal component analysis is adopted to reduce the feature dimension and the complexity of data storage and computation. In the cluster level, for the security of sensors, normal sensor nodes employ the transductive confidence machines for K-nearest neighbors for anomaly detection, and cluster heads use the support vector machine based on particle swarm parameters optimization to classify the misuse detections for the detected anomaly data. In the base station level, for the security of cluster heads, the anomaly detection and misuse detection technologies are combined to deal with the moni- toring data delivered by cluster heads, which improves the detection probability while preserving low false alarm probability. Simulation results show that the proposed detection algorithm can improve the accuracy of detection even in the case of small samples.
出处
《信息与控制》
CSCD
北大核心
2013年第6期670-676,共7页
Information and Control
基金
国家自然科学基金资助项目(61373126)
江苏省自然科学基金资助项目(BK20131107)
关键词
无线传感器网络
分级入侵检测
K近邻直推式信度机
粒子群优化
支持向量机
wireless sensor network
hierarchical intrusion detection
transductive confidence machine for K-nearest neighbors
particle swarm optimization
support vector machine
