面向链路比特流的未知帧关联分析  

Data Link Bit Stream Oriented Association Analysis on Unknown Frame

作　　者：薛开平[1] 柳彬[1] 王劲松[2] 李威[1] 薛颖杰 XUE Kaiping LIU Bin WANG Jinsong LI Wei XUE Yingjie(School of Information Science and Technology, University of Science and Technology of China, Hefei 230027, China Southwest Electronics and Telecommunication Technology Research Institute, Chengdu 610041, China)

机构地区：[1]中国科学技术大学信息科学技术学院,合肥230027 [2]西南电子电信技术研究所,成都610041

出　　处：《电子与信息学报》2017年第2期374-380,共7页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(61379129);中国科学院青年创新促进会人才基金(2016394)~~

摘　　要：在电子对抗中,截获到对方的通信比特流序列之后,当链路协议类型未知时,现有的协议解析工具往往无法分析比特流所承载的有用信息。为了获取比特流承载信息,首先需要切分比特流得到链路帧。该文根据链路帧结构的一般规律,提出一种基于数据挖掘的比特流切分算法。通过频繁序列统计、关联规则分析以及关联规则整合,识别出比特流中标识帧起始的多重关联规则序列。测试结果表明,该算法能够从未知比特流中提取有效的切分标识,正确实现比特流切分。与同类基于数据挖掘的比特流分析方法相比,该算法复杂度低,输出结果唯一且可信度高。In the electronic countermeasure, the opponent＇s bit stream can be captured. However, without any knowledge about the type of data link protocol, the existing protocol analyzing tools can not analyze the useful information from the bit stream. To further get the carried information, the bit stream should be segmented to frames firstly. According to the general rules of frame structure, a bit stream segmentation algorithm is proposed based on data mining, in which, the multi-association rule indicating the beginning of frames can be identified by using frequent sequence statistics, association analysis and association rules integration. The test results show that, this algorithm can extract the valid segmentation flag from unknown bit stream and segment the bit stream correctly. Compared to the similar data mining based bit stream analyzing algorithms, this algorithm can be more efficient and produce a unique result which is of high reliability.

关 键 词：链路比特流 未知帧 频繁统计 关联分析 切分 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]