【作者】 张英杰；

【导师】 高强；

【作者基本信息】 吉林大学 ， 计算机应用技术， 2011， 硕士

【摘要】 随着无线网络技术和车辆电子的结合,车载网络与信息系统成为现代智能汽车的重要组成部分。汽车已经不再是人们单纯的代步工具,许多人致力于将自动控制、汽车电子、多媒体、互联网、无线通信等多种技术融合于车载网络。随着车载系统的日趋成熟,各种服务(车载电子导航、交通优化、无线上网、支付服务)的问世,已经给人们的生活带来了很大的便利,但这些服务应用的同时也不可避免地牵涉到用户的个人数据和隐私的安全问题,因此车载网络的安全问题成为目前的焦点。随着我国经济的快速发展,目前国内拥有汽车的用户已经超过一千万人而且这个数目还在持续不断地上升。各大汽车制造厂商为了在竞争中获得最大的利益都纷纷致力于车载网络的安全性研究,车载网络有望实现更成熟可信的无线自组织通信系统。可信计算的诞生为车载网络的发展注入了新活力。可信计算的主要目的是通过终端的安全性来保证整个网络的安全,因为终端是产生网络数据的主要来源,同时还是与网络中服务器或是其他节点通信的重要组成部分,可见确保终端的可信性尤为重要。TCG小组提出的TNC架构,在一定程度上保证了终端的可信性。在TNC架构中的身份验证和完整性信息检测双重验证使得某些恶意入侵的节点(冒充身份用户)被屏蔽,得不到接入网络的许可。本文介绍了车载自组织网络的背景及其特点,还详细地介绍了TNC架构及其通信原理,并在现有的可信网络研究的基础上提出了基于可信网络接入技术的通信机制。可信接入技术主要通过对终端用户的身份认证和完整性信息认证来对终端实施访问控制,网络内部的控制主要通过4A框架下的审计来实现。本文还提出了在4A框架下的一种应用于安全审计的规则提取模型,也就是本文中详细介绍的决策树分类算法模型。本文的决策树分类算法在原有的经典的ID3决策树算法的基础上做了改进,以适用本文的车载网络环境的分类算法模型。利用该决策树算法模型对网络数据进行挖掘,以用来对车载网络中产生的攻击行为的数据进行统计分类,通过建立决策树模型对这些行为进行特征提取。最终可以将这些特征(实际上就是规则)放入到审计的动态规则库中。TNC架构中对终端的控制体现在两个方面,一方面是对终端用户的访问进行控制,另一方面是对接入网络的用户行为的审计。随着车载终端用户的改变以及时间的推移,审计数据最终会越来越大,最后形成一个庞大的数据平台,利用这个数据平台可以为用户提供安全服务。安全审计是对已经接入网络的用户的上网行为进行记录和监控,是网络安全体系中的重要一环,它在一定程度上弥补了防火墙和入侵检测的不足。本文结合车载网络数据的特点对传统的决策树模型进行修改,提出应用于安全审计的决策树新模型,同时经过多次实验验证,结果表明决策树分类模型能够适应车载网络数据的特点能有效地防止恶意攻击的发生。随着车载网络中数据的变化,决策树模型也随之发生改变,最后提取出的规则也会随着时间的改变而改变,因此提取出的规则最后要放入到动态规则库,用户可以不定期地对该规则库进行更新,以保证数据的及时有效性。更多还原

【Abstract】 As the combination of wireless network technology and vehicle electronics, vehicle network and information system become the important part of the modern intelligent vehicle. Vehicle is no longer simply a means of transport for people, many people are committing to the automatic control, automotive electronics, multimedia. Internet, wireless communications and other technologies into vehicle network. As the vehicle system matures, the advent of kinds of services (such as vehicle electronic navigation, traffic optimization, wireless internet access, payment services)and so on, has brought a great convenience to the life of people. However, the application of these services also inevitably involves the user’s personal data and privacy issues, therefore the safety of vehicle networks become the focus. With the rapid economic development of our country, the owners of car have more than ten million in domestic and the number is continually rising. Major automobile manufacturers have committed to the safety of vehicle networks in order to obtain the greatest benefit in the competition, it is expected that the vehicle networks will achieve a more mature and trusted wireless communication system.The birth of trusted computing injected new vitality to the development of vehicle network. The main purpose of trusted computing is to ensure the security of the entire network through the safety of the terminal, because the terminal is a major source which can generate data of network, and it is an important part of communicating with the server or the other node in the network. Ensuring the credibility of the terminal is particularly important. The TNC architecture proposed by the TCG’s group ensures the credibility of the terminal to a certain extent. In the architecture of TNC, both the authentication and the detection of integrity of information make certain malicious invasion node (posing as the user) be shielded without the permission of accessing to the network.This article describes the background of vehicle self-organizing network and its features, and the description of the architecture of TNC and the communication theory in detail. As well as research in the existing trusted network is proposed based on trusted network connected technology based on the mechanism of communication. At the same time, a rule extraction model is proposed in the part of audit under the framework of 4A, which is applied in the safety audit, that is the decision tree model which is described in detail.In this paper, the decision tree is used to mine data of network, generate on-board network attack data for statistical records, extract feature of these acts through the establishment of decision tree model.The control of architecture of TNC is reflected in two aspects to the terminal, one is to control access of network for users, the other is the audit of user’s behavior online. As the end-users of vehicles changing and time passing, audit data is increasing ultimately and finally a large data platform is formed which is using the trusted platform system providing users with security services.Safety audit is recording and monitoring the behavior of the users who are already admitted to access to the network, therefore it’s an important part of network security system, and it also makes up for lacking of firewall and intrusion detection to some extent. In this paper, a new decision tree model is proposed for the security audit, it is combined with the characteristics of data of network, and it is modified on base of the traditional decision tree. After many experiments, the results show that the decision tree classification model can adapt to the characteristic of data of vehicle network and it also can prevent malicious attacks effectively. With the changes in the data of network, decision tree model will also be changed with it, so the rules extracted from the database will be put into a dynamic database, users can update the rule base from time to time to ensure the validity of data in a timely manner.更多还原