【作者】 王泽；

【导师】 陈永乐；

【作者基本信息】 太原理工大学 ， 计算机技术， 2019， 硕士

【摘要】 随着现代通信及计算机技术的发展,依据名字、相貌、音色甚至身高等因素进行真实生活的身份认证,也可以为虚拟的网络空间提供安全、稳定、“真实”的身份验证。伴随基于所知、所是和所有的口令、实物设备和生物特征的单因子及多因子认证中口令遭受字典攻击、智能卡携带不便及丢失、生物特征隐私性强和缓慢变化等问题的严重,对现代信息社会造成巨大威胁。因此2009年美密会上,N.Chandran,V.Goyal,R.Moriarty和R.Ostrovsky共同提出使用地理位置信息验证策略,可以很好地解决主动认证中信息泄露造成的非法访问等安全隐患。无线网络的普及,为利用无线信号的变化描述复杂室内环境的位置和人体行为特征提供了便捷条件。基于信号接收强度(RSSI)的无线信号特征虽然可以提供室内位置信息作为辅助认证,但存在较大波动及误差。随着OFDM技术和CSITOOL设备的普及,可以获取物理层各个子载波连续、空时唯一的CSI(信道状态信息)为复杂室内环境位置提供更加细致、多样的感知描述。CSI是对各个子载波在物理层面的描述,因此在获取、比较中只涉及底层硬件操作,在降低数据复杂度提高后期数据处理、认证的同时简化环境搭建有效降低实验成本,为利用CSI进行基于室内位置和行为动作的辅助识别提供基础。本文以无线信号的CSI特征为研究对象,从基于CSI位置指纹和动作识别两个方式入手,提供基于CSI特征的辅助身份认证技术,并在数据处理和匹配方引入新方法,具体工作如下:1、总结身份认证原理,列出主流身份认证技术研究内容,在对比分析双因素、三因素认证基础上,提出使用基于CSI数据特征作为辅助认证,结合其他主流认证方式的多因素、被动式的持续认证方案。2、介绍并对比描述无线信号特征的两大属性:RSSI和CSI。在对RSSI和CSI概念及数据处理技术介绍后对比信号稳定和敏感性,物理层CSI信息可以有效地降低室内多径效应的干扰,从而提高基于位置和行为特征辅助认证方案的准确率。3、总结和对比基于CSI室内定位研究现状,在数据处理和特征提取中采用改进的K-means算法,从选取初始聚类点和后期迭代计算中提出改进,增强聚类效果,提高算法时效性。4、总结和对比基于CSI人体行为感知研究现状,本文采用低通滤波去噪,对去噪后数据进行PCA降维处理,提出主要描述行为特征的信道状态信息组成低维CSI动作特征,利用一对一多分类SVM(support vector machine)对动作CSI特征进行分类匹配,提高对低维度CSI动作数据分类识别,增强实效性。5、构建基于CSI的认证地图,利用CSI位置和动作特征对访问WLAN的用户进行N:1的CSI数据验证,并在基于CSI位置特征辅助认证中,对访问认证失败的用户进行定位追踪。综上所述,本文提出利用基于CSI位置和固定动作特征作为WLAN访问的辅助认证方法,并结合其他主流网络身份认证技术,可以提供一个多层次、被动式和安全便捷认证服务。本方案要求资源访问者必须在固定的位置使用固定的设备或者做出要求动作进行数据匹配才可以完成WLAN认证,访问资源。因此,本方案可以规避远程、异地登录等问题,并在复杂室内环境下,针对不同工位提供小范围的权限认证服务和不同用户鉴别身份认证服务。最后,对本文的工作进行全面总结和阐述,提出基于CSI特征的辅助身份认证技术今后还需努力的工作和方向。更多还原

【Abstract】 With the development of modern communication and computer technology,identity authentication in real life based on name,appearance,timbre and even height can provide security,stability and "real" identity verification for virtual cyberspace.With the serious problems of dictionary salary,inconvenience and loss of smart cards,strong privacy and slow change of biological characteristics,passwords in single-factor and multi-factor authentication based on knowledge,ownership and all passwords,physical equipment and biological characteristics pose a great threat to the modern information society.Therefore,at the 2009 US Secret Conference,N.Chandran,V.Goyal,R.Moriarty and R.Ostrovsky jointly proposed the use of geographic location information verification strategy,which can solve the security risks such as illegal access caused by information leakage in active authentication.The popularity of wireless networks provides a convenient way to describe the location and human behavior of complex indoor environments using changes in wireless signals.The wireless signal characteristics based on signal received strength(RSSI)can provide indoor location information as auxiliary authentication,but there are large fluctuations and errors.As the physical layer description of the wireless network describes the channel state information(CSI),the space-time uniqueness of the wireless channel response is utilized by the channel state information(CSI)physical layer authentication technique,by comparing the correlation between successive channel responses.Perform message authentication.Since the comparison of continuous channel responses involves only hardware operations,physical layer authentication has the characteristics of fast and efficient,and is very suitable for continuous and realtime authentication of terminals with limited resources and resources in wireless devices.This thesis takes the application of physical layer authentication technology in network identity authentication as the research object,and mainly accomplishes the following work:1.Firstly study the current application of identity authentication technology,analyze the advantages and disadvantages of various identity authentication technologies,and propose the use of CSI-based location and action features as a network authentication assistance method,combined with other identity authentication methods to provide a multi-level,passive and secure Certification scheme.2.A CSI-based WLAN-assisted identity authentication technology is proposed.Orthogonal Frequency Division Multiplexing(OFDM)is used to obtain fine-grained channel state information(CSI)to describe the signal characteristics of position and fixed motion,effectively reducing indoor multipath effects.The interference improves the accuracy of location fingerprint based authentication and attack detection and location.3.Summarization and comparison Based on the current research status of CSI indoor positioning,the improved K-means algorithm is adopted in data processing and feature extraction.The improvement is proposed from the selection of initial clustering points and later iterative calculations,the clustering effect is enhanced,and the timeliness of the algorithm is improved.4.Summarizing and comparing the current research status of human behavior perception based on CSI,in the processing of continuous changing feature data,this thesis adopts low-pass filtering to denoise,and performs PCA dimensionality reduction on denoised data,and proposes channel state information mainly describing behavior characteristics.The low-dimensional CSI action feature uses a one-to-one multi-class SVM(support vector machine)to classify and match the action CSI features,improve the classification and recognition of low-dimensional CSI action data,and enhance the effectiveness.5.The CSI-based authentication map is constructed,and the NSI access authentication is performed on the user accessing the WLAN by using the CSI location and the action feature,and the attack detection and location are performed on the user who fails the authentication.In summary,this thesis proposes to use CSI location and fixed action features as the auxiliary authentication method for WLAN access,and combined with other mainstream network identity authentication technologies,it can provide a multi-level,passive and secure and convenient authentication service.This solution requires resource visitors to use fixed devices in a fixed location or make required actions for data matching to complete WLAN authentication and access resources.Strong authentication services can be provided to the core area on a smaller scale to ensure information security.Finally,the thesis summarizes and expounds the work of this thesis,and proposes the work and direction of the auxiliary identity authentication technology based on CSI features in the future.更多还原