基于小波分析的无线传感网实时异常检测算法  

Wavelet Analysis-Based Real-Time Anomaly Detection Algorithm for Wireless Sensor Network

作　　者：李致远[1] 朱求志 吴永焜 唐振宇[1] 胡华明[1] 

机构地区：[1]江苏大学计算机科学与通信工程学院,江苏镇江212013

出　　处：《南京师大学报（自然科学版）》2014年第1期87-92,共6页Journal of Nanjing Normal University(Natural Science Edition)

基　　金：国家自然科学基金(61202474;61103195);江苏省自然科学基金(BK20130528);江苏大学高级专业人才科研启动基金项目(12JDG049);江苏大学本科生创新计划项目(2012075)

摘　　要：异常检测技术能够检测到未知攻击,对于保障无线传感器网络安全具有重要意义.当前的异常检测技术实时性差,误报率高且计算量大,因此,无法直接应用在无线传感器网络中.鉴于此,提出基于小波分析的实时无线传感网异常检测(Wavelet Analysis-Based Real-time Anomaly Detection,WARAD)算法.在整个检测过程中,WARAD算法采用了逆向获取实时网络流量,然后通过对小尺度区间使用小波系数方差法计算Hurst值,从而提高异常检测的实时性、准确率,并降低求解Hurst值的运算复杂度.最后,在MeshIDE平台上实现了基于WARAD算法的异常检测系统,实验结果表明此算法极大地提高了无线传感网环境下异常检测的实时性,并降低了异常检测的误报率和漏报率.Anomaly detection can detect new and unknown attacks,which has great significance on the wireless sensor networks security. Nowadays,the proposed anomaly detection schemes has poor real-time,high false positive rate and the large amount of computational overhead, and hence the schemes are not suitable for wireless sensor networks. In this paper,a wavelet analysis-based real-time anomaly detection （ Wavelet Analysis-based Real-time Anomaly Detection, WARAD） algorithm for wireless sensor network is proposed. Throughout the detecting process, the WARAD algorithm reversely collects the real-time network traffic,and then uses the variance of the wavelet coefficients in the small-scale interval to compute the Hurst values,which can improve the real-time and the accuracy of anomaly detection,and reduce the computational complexity of solving the Hurst values. Finally,the WARAD algorithm-based intrusion detection system is implemented on the platform of MeshIDE. The experimental results showed that the proposed algorithm greatly improved the real-time of anomaly detection for wireless sensor networks,and reduced the false positive rate and the false negative rate of anomaly detection.

关 键 词：无线传感器网络 安全 异常检测 小波分析 HURST参数 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]