基于规划识别的入侵检测研究

Research on the Intrusion Detection Based on Plan Recognition

规划识别是人工智能的重要研究分支之一,在入侵检测领域中已有初步的应用。本文在介绍规划识别和入侵检测基本概念的基础上,按照规划识别方法分门别类地研究了基于事件层的规划识别、基于贝叶斯网络的规划识别、基于扩展目标规划图的规划识别、彩色Petri网、对手规划、行为状态图等在入侵检测领域的应用现状和进展;接着深入分析了规划识别和入侵检测的关系和相似之处;最后讨论了基于规划识别的入侵检测存在的问题,并指出了未来的发展趋势。本文综述了智能规划在入侵检测中应用的关键技术和存在的问题,研究内容对于相关人员从事入侵检测研究具有重要的参考价值。

Plan recognition is one of the important branches of artificial intelligence,and it has been preliminarily applied in intrusion detection systems（IDSs）.The fundamental conception of plan recognition and IDS are introduced firstly.And then it studies on the applications and progresses of plan recognition in IDS according to the classification of plan recognizing methods,such as plan recognition based on event,plan recognition based on the Bayesian network,plan recognition based on the extended goal graph,colored petri net,adversarial planning and action state diagram.Subsequently,the common characteristics of plan recognition and IDS are thorough analyzed.Finally,the development trends and the existing problems of this domain are discussed.This review deals with the key technology and the existing problems of plan recognition implicating in IDS,which is very helpful for the related personnel research in the logistics fields.