【作者】 王鑫；

【导师】 贾庆轩；

【作者基本信息】 北京邮电大学 ， 机械电子工程， 2015， 博士

【摘要】 物联网作为当今信息产业发展的第三次革命浪潮,是全球信息产业革命新一轮竞争的制高点,从而以物联网为核心的各项信息技术革命成为了世界各国在信息领域的国家战略,得到了迅猛发展。无线射频识别(RFID)是物联网架构下最为关键的自动识别技术,其可以感知物件信息并为顶层应用提供及时快速的信息源,作为物联网感知层神经末梢将物联网延伸到万事万物。由于RFID技术具有可读写、远距离识别及批量读取等诸多优点,目前该技术己逐渐取代条形码广泛应用于交通、物流、生产制造及航空等自动识别领域。然而,随着RFID技术的广泛使用,标签的造价势必越来越低以适应市场需求,低廉的造价致使标签具有存储空间低及计算能力弱的不足。同时,对RFID系统识别效率及无线信道安全与隐私的防护能力提出了更高的要求。因此,如何在存储空间低及计算能力较弱的被动标签基础上,提出高识别效率的防碰撞算法及强安全性的安全机制,是目前RFID领域亟需解决的关键问题,对于推进RFID系统进一步大规模推广具有极其重要的意义。本文分别围绕一般应用环境及特定应用环境下的RFID防碰撞算法,以及具有后台服务器下的超轻量级安全认证协议和移动场景下的安全搜索协议开展了深入研究,主要研究内容包括以下几个方面：(1)基于位图构建的RFID自适应N树防碰撞算法研究。针对一般应用环境下批量标签通过共享信道多路访问而导致的碰撞问题,提出两种基于位图构建的RFID自适应N树防碰撞算法,即单帧自适应位图构建N树防碰撞算法(SFBTNA)及多帧自适应位图构建N树防碰撞算法(MFBTNA)。首先,利用极大似然位跟踪标签数估算法(MLBE)估算标签数量。其次,在MLBE基础上提出了探针法构建位图的方法,SFBTNA算法根据位图在单个识别帧内自适应调整最优N值并构建N树,而MFBTNA算法则逐帧调整最优N值并构建多个N树。最后,两种算法均采用跟踪树算法(CT)对N树下的碰撞子树进行识别,位图构建与CT算法的结合使用消除了空闲时隙,并极大地降低了碰撞时隙的影响。因此,有效提高了RFID系统的识别效率。文中分别推导了SFBTNA算法及MFBTNA算法的最优帧长取值区间,使得系统获得了较高的识别效率；(2)特定应用环境下自适应型分组N叉树RFID防碰撞算法研究。针对读写器重复识别滞留标签的特定应用环境,提出了自适应分组N叉树防碰撞算法(ACBGN)。首先,利用阻塞机制对滞留标签和新到标签进行划分,并分阶段识别滞留标签和新到标签,以避免两类标签之间的碰撞。其次,分别根据滞留标签和新到标签的识别特点提出了对应的识别算法,旨在分别提高对两类标签的识别效率。对新到标签的识别采用分组和多叉识别两种策略相结合的方法,降低了标签之间的碰撞概率,文中对不同分叉数时的最优分组系数进行了推导,获得了最优分组系数及分叉数的组合,有效提高了系统对新到标签的识别效率。同时,提出了多Hash映射多帧耦合型签到算法识别滞留标签,其采用单标签签到识别及两个标签耦合识别的策略减少对滞留标签的识别耗时,并结合位图设置Hash优先级以压缩映射位图,文中推导了最优Hash映射次数及帧长设置,充分利用了滞留标签的ID信息,从而有效降低了系统对滞留标签的识别耗时。因此,在对新到标签和滞留标签识别效率分别提高的基础上,ACBGN算法的识别效率得到了较大提高；(3)超轻量级安全协议的设计与分析。在具有后台服务器情况下,通过对现有超轻量级安全协议设计中常见的漏洞进行分析,提出了两个超轻量级安全协议,即抗去同步化超轻量级安全认证协议(URP)和超轻量级安全切换协议(UTFP)。URP协议通过在交互消息中加入新鲜分散因子,并采用新旧动态ID更新的方式抵抗去同步攻击及重放攻击。同时,结合加密因子MixBit和Per的加密特性以达到更好的混淆与扩散效果,使得URP协议可以抵抗泄漏攻击。最后,其仅利用对计算性能要求较低的位运算来实现读写器与标签之间的无线链路安全防护,对标签的存储能力和计算能力要求较低。文中利用基于攻击检测的非形式化分析方法对提出的URP协议进行了安全性证明,保证了URP协议满足完整性及保密性,并可以有效抵抗无线信道中的常见攻击方式。UTFP协议在URP协议基础上,利用被动标签可生成伪随机数的特性对失去同步性的URP协议进行再同步。与此同时,在协议遭受主动攻击时,UTFP协议自动启动再同步机制更新动态密钥与动态ID,从而使得协议获得更高的安全性及隐私性防护。(4)可证明安全的无服务器型安全搜索协议研究。针对无后台服务器情况下的移动读写应用场景,提出了一个轻量级无服务器型安全搜索协议(PLSP)。首先,通过严格时间戳的方式对移动读写器时间权限进行管理,提升了移动读写器安全性。其次,根据实际应用环境下的标签数,利用Hash碰撞的思想实现了可调节的隐私强度和搜索效率。最后,利用松散时间戳抵抗重放攻击,其具有适应低价标签计算性能的特性。与此同时,在UC框架下证明了提出的安全搜索协议满足并发安全性,并在标准安全模型下证明了协议可有效抵抗重放、去同步、跟踪及匿名性等攻击方式。更多还原

【Abstract】 The internet of things (IoT), as the third revolution in information technology wave, is the new competition commanding heights of the global information industry. And the IoT-focused information technology revolution has become national strategy in the field of information. Radio-frequency identification (RFID), as the IoT perception layer tentacles providing information source for the top-level application through identifying and tracking tags attached to objects, is the most key automatic identification technology in IoT framework. Therefore, RFID is widely used in automatic identification field such as transportation, logistics, aviation, and gradually replaced the bar codes base on the advantages of read-write, long-distance identification and batch read. However, the decreased tag cost, increased performance to identification efficiency and radio channel security privacy protection drive a widespread in RFID usage by market demand. Hence, how to propose the highly effective anti-collision algorithm and the strong safety security mechanism became the key problems in the field of RFID under passive tags of the weak storage capacity and computing power, which is very useful to promote RFID system widely used. This paper studied on the RFID anti-collision algorithm under the environment of unknown and specific application, the ultra-lightweight security authentication protocol based on having backend-server, and the security search protocol in the movement environment. The following key works are mainly researched:(1) Research on RFID adaptive N tree anti-collision algorithm based on bitmap construction. This chapter put forward two RFID adaptive N tree anti-collision algorithms based on bitmap construction (SFBTNA and MFBTNA) for the RFID collision problem caused by multiple access of shared channel under unknown environment. First of all, it estimated the number of tags effectively by the maximum likelihood tag numbers estimation method (MLBE).. Secondly, it proposed that the probe method constructed bitmap, and SFBTNA adaptively adjusted optimization value of N and built N trees through bitmap under one frame based on MLBE. MFBTNA adaptive adjusted optimization value of N and built N trees via frame-by-frame. Further, SFBTNA and MFBTNA identified the collision sub-trees using tracking tree (CT) algorithm. The combination of bitmap and CT eliminated idle time slot and greatly reduced the impact of collision time slot, and then greatly improved the identification efficiency of RFID system. This study derived the optimal frame length of SFBTNA and MFBTNA, and built the optimal N tree by the bitmap, and obtained the optimal throughput, respectively.(2) Research on adaptive grouping N-ray tree of RFID anti-collision algorithm under specific application environment. This chapter proposed adaptive grouping N-ray anti-collision algorithm (ACBGN) algorithm for some readers repeatedly identifying staying tags under the specific application environment. First of all, we divided the staying tags and arriving tags by the blocking mechanism, and then identified them in different stages for avoiding the collision between two kinds of tags. Secondly, we put forward the corresponding identification algorithm for improving identification efficiency of staying tags and arriving tags based on the identification feature of them, respectively. The arriving tags were identified under combining grouping and N-ray identify for reducing the collision probability between tags. Further, the idle slots were reduced through collision bit tracking in order to improve the throughput of the arriving tags identification. This paper derived the optimal number of groups under different rays, and obtained the optimal combination. The staying tags were identified by the multi-frame coupling algorithm for getting the optimal identification efficiency. The algorithm compressed mapping bitmap through identification strategy of single-tag sign in, two-tags coupling, and setting the Hash priority. This paper derived the optimal number of Hash mapping and improved the identification efficiency. The ACBGN algorithm obtained high identification efficiency and and was suitable for rapid automatic identification in the specific application scenario field based on high staying tags and arriving tags identification efficiency.(3) Research on the ultra-lightweight security protocol design and analysis. This chapter put forward two ultra-lightweight security protocols, URP (the ultra-lightweight security authentication protocol of resist on desynchronization attacks) and UTFP (the ultra-lightweight security transfer protocol), through analyzing common vulnerabilities in the existing ultra-lightweight security protocols in case with backend server. URP protocol resisted against desynchronization and replay attacks by adding new synchronization factor in interactive message and combining old and new dynamic ID updated ways. URP protocol could resist leak attack via combining the encryption features of MixBits and Per for better confusion effect. It protected the wireless link security between reader and tag only through simple bit operation. This paper verified that our proposed URP protocol satisfied the integrity and confidentiality of security features and could resist common attacks of radio channel. UTFP protocol could resynchronize the desynchronized URP protocol by generating pseudo-random number of passive tag. Also the protocol initiated resynchronization mechanism to update dynamic key and ID and then obtained higher security and privacy under the protocol suffer active attacks.(4) Research on provable secure service-less security search protocol. This chapter proposed a lightweight service-less security search protocol for the mobile reader scenario under the situation of without the background server. First of all, the protocol achieved the time rights management by strict time stamp, and then improved the security of mobile reader. Secondly, the protocol realized adjustable privacy intensity and search efficiency by the number of tags in the practical application environment based on Hash collision. Finally, the protocol adapted computing performance of low-cost tags via loose timestamp style, and could resist against replay attacks effectively. This paper verified that the protocol satisfied concurrent security under UC framework, and could effectively resist against replay, desynchronize, tracking and anonymity attacks in standard security model.更多还原