基于SVM的MANET路由层入侵检测*

针对MANET（移动自组织网络）路由层的攻击，通过对MANET路由层AODV（Ad hoc on-demand distance vector）路由协议交互行为的分析，提取了9个路由交互过程特征，将入侵检测问题转换为对正常行为和异常行为分类识别问题，采用SVM（支持向量机）算法，设计了一种分布式异常检测系统。仿真结果表明，使用该检测系统的检测率达到97%以上，从而验证了该系统的可行性，同时也验证了所提取的路由交互行为特征的有效性。

Intrusion detection for routing layer of MANET based on SVM

Aiming at the attack of routing layer of MANET through analyzing the features of the traffic behavior of the AODV routing protocol in routing layer of MANET, this paper picked up nine features of the traffic behavior and translated the problem of intrusion detection to classifying the behavior between normal and abnormal.As a result, designed a distributed SVM-besed anomaly detection system.The simulations show that the accurate recognition rate of this system is above 97%, and which demonstrates that the system is effective and the features of the traffic behavior which this paper picked up is feasible.