利用RFID中间件的特点,加入入侵检测异常检测模块,通过对入侵者的攻击数据流与RFID系统内部的数据流进行对比后提取特征向量,并通过对入侵数据流对应子模块的归纳建立有限自动机,对有限自动机的归并来提高系统入侵检测的效率。此外,参照归并后的有限自动机对攻击进行分类,通过对攻击本质的分析与提取来检测一部分入侵检测自动机里没有相应的攻击。最后,对固定攻击比例的访问事件样本乱序进行试验,试验结果表示系统在3轮测试中均得到稳定的检测率。
Based on the characteristics of RIFD middleware, the paper adds an intrusion detection system in RFID middleware to gather the feature vector from the comparison between intrusion data stream with RIFD internal data stream and establish finite automation according to intrusion data steam corresponding to sub-module. The efficiency of the intrusion detection system is improved via the consolidation of some of finite automation. Besides, the proposed system can achieve the classification of RFID attacks from finite automation, and even detect some unknown attacks through analyzing the nature of the attacks. By randomizing the order of access events which has mixed attack events in fixed proportion, the system can achieve stable detection rates in 3 rounds tests.