ISSN: 1673-825X    Imprint: Chongqing University of Posts and Telecommunications Journal
Volume listings
  • Current volume
  •  
  • Search by keywords:
  • Search by 
  • Keywords
  • To
  • Journal archive
Related Links
物联网终端智能识别系统设计与实现
Design and implementation of intelligent identification system for IoT terminals
DOI:10.3979/j.issn.1673-825X.2019.04.003
Received:January 02, 2019  Revised:June 17, 2019
中文关键词:物联网(IoT)  终端智能识别  设备指纹  机器学习算法  分类器
英文关键词:internet of things(IoT)  terminal intelligent identification  device fingerprinting  machine learning algorithm  classifier
基金项目:国家自然科学基金(61872447);重庆市自然科学基金(CSTC2018JCYJA1879);重庆市教委科学技术研究重点项目(KJZD-K201802101)
Author NameAffiliationE-mail
ZHANG Li Rongzhi College of Chongqing Technology and Business University, Chongqing 401320, P. R. China zhangli_zju@163.com 
WANG Zuoyue Key Laboratory of Computer Vision and System, Ministry of Education, Tianjin University of Technology, Tianjin 300384, P. R. China 568365250@qq.com 
WANG Chundong Key Laboratory of Computer Vision and System, Ministry of Education, Tianjin University of Technology, Tianjin 300384, P. R. China michael3769@163.com 
MA Yunfei Logistics Planning Bureau, Army Department of PLA, Beijing 100043, P. R. China 61298688@qq.com 
XIANG Chaocan College of Computer Science, Chongqing University, Chongqing 400044, P. R. China chaocan@gmail.com 
Hits: 34
Download times: 15
中文摘要:
      终端智能识别是物联网应用的关键技术,是物联网安全体系构建的基础。针对物联网终端智能识别问题,建立了一种以设备指纹为动态特征标识的物联网终端智能识别实验系统。该系统由终端检测模块、模型训练模块以及智能识别模块构成,其中,终端检测模块利用Nmap工具扫描并自动采集设备指纹;模型训练模块分别利用决策树、逻辑回归与朴素贝叶斯等机器学习算法训练分类器;智能识别模块接收识别任务并调用前2模块完成设备指纹采集与分类识别处理。实验结果表明,决策树分类器在整体数据集上的平均识别率为98.1%,对于是否是物联网设备的判断识别率为98.7%,对于具体设备类型的识别率为98.2%,均保持较高识别水准,且优于其余2种算法识别器。因此,采用设备指纹与决策树算法结合识别物联网设备是可行的。
英文摘要:
      Intelligent terminal identification is a key technology of internet of things (IoT), and is also the foundation of building the security system of IoT. Aiming at the problem of intelligent identification of IoT terminals, an experimental system is established based on dynamic device fingerprint. The system mainly includes terminal detection module, model training module and intelligent identification module. Terminal detection module scans IoT terminals using Nmap software tool and collects fingerprints of devices automatically. The model training module uses decision tree, logical regression and naive Bayesian to train the classifiers separately. The intelligent recognition module receives recognition tasks and calls the first two modules to complete the fingerprint acquisition and classification. The experimental results show that the average identification rate of the decision tree classifier is 98.1% in the whole data set, 98.7% for IoT device confirmation and 98.2% for the device type. Therefore, it is feasible to combine the device fingerprinting and decision tree algorithm to identify IoT terminals.
HTML    View PDF   View/Add Comment  Download reader