In the era of rapid development of Internet technology,security has always been the focus of attention all over the *** authentication plays an important role in protecting the security of computer *** identity authen...
详细信息
In the era of rapid development of Internet technology,security has always been the focus of attention all over the *** authentication plays an important role in protecting the security of computer *** identity authentication methods usually use account and password to protect the security of computer system,but this method can not provide continuous *** continuous protection can be realized through the identity authentication of the user and the human-computer interaction data of the computer *** most important human-computer interaction data is the mouse and keyboard *** that the mouse data contains less user privacy and is relatively safer,it has attracted extensive attention.
Mouse behavior authentication is divided into static authentication and persistent *** with static authentication,persistent authentication has fewer restrictions on the user's use of the mouse,and can continuously monitor the user's identity,so it has become a research ***,the existing identity authentication methods have the problems of low authentication accuracy and long authentication *** thesis studies these issues,and the specific research contents are as follows:
1.A novel method for continuous identity authentication based on mouse behavior is *** method first divides the mouse event sequence into various mouse behaviors at different levels,and mines mouse behavior features from various aspects based on mouse behavior to improve the authentication *** feature selection algorithm removes redundant features to reduce model complexity and modeling ***,a binary classification algorithm is used to train the authentication *** experimental results show that this method can obtain higher authentication accuracy with fewer features extracted.
2.A continuous identity authentication method based on fusion trust model is *** order to reduce the authenticatio
随着汽车数量的迅速增长,交通安全和交通拥堵等问题日益严重。车联网是实现车辆与周围的人、车、交通基础设施和网络等全方位连接和通信的新一代智能交通技术,被认为是解决上述问题的关键。由于车联网具有网络拓扑结构变化频繁、通信场景多样化、节点密度较大等特点,使得车联网中的车辆节点比传统的网络节点面临更多、更复杂的网络内部攻击问题。尤其当恶意车辆故意广播与交通安全相关的虚假消息时,车辆的驾驶决策会受到很大影响,此类恶意行为会严重影响到交通系统的安全和效率。信任管理机制能够对车联网中车辆的可信度进行有效度量,帮助车辆选择可信的车辆进行合作。针对车联网中的信任管理问题,本文主要做了以下研究:(1)本文提出了一种用于评估车辆可信度的信任模型,采用基于贝叶斯信任推断和TrustRank算法分别评估车辆的局部信任值和综合信任值。计算局部信任值时,在当前贝叶斯信任理论研究的基础上加入惩罚因子和调节函数来降低开关攻击对评估结果的影响,并引入时间滑动窗口对局部信任值进行更新;然后根据局部信任值构建所有车辆的信任链接图;最后综合信任链接图中车辆输入和输出的信任关系通过TrustRank算法来计算车辆的综合信任值。(2)本文提出了基于区块链技术的分布式车联网信任管理方案,通过将RSU连接成一个分布式的区块链网络实现对车辆信任数据的存储和管理;其次改进了PoA(Proof of Authority,PoA)共识算法来提高信任数据更新的效率;研究了区块链技术和以太坊平台,构建了一个基于以太坊平台的去中心化应用框架,并依据该框架在以太坊平台上开发了去中心化的车联网信任管理应用,最后对该应用的安全性能进行了分析和对比。使用OMNeT++5.5.1、Veins5.0和SUMO1.2.0仿真平台对本文提出的模型进行验证,结果表明,加入惩罚因子和调节函数的局部信任值计算模型有利于抵抗开关攻击;本文所提出的信任模型与其它相关模型相比,真阳性率和真阴性率都有明显提高,从而验证了模型的有效性。最后基于Ethereum平台完成了智能合约的开发部署和以太坊开发环境的搭建,利用truffle、Ganache等工具及编程语言实现了本文所提出的信任管理方案,从系统框架、功能设计、设计流程等方面详细介绍了方案实现的过程,并对方案进行了测试,验证了所提出方案的可行性。
暂无评论