无线传感器网络的密钥分配算法及其在路由安全协议中的应用研究

发布时间：2018-04-23 11:51

本文选题：无线传感器网络 + 密钥分配算法　；参考：《电子科技大学》2014年硕士论文

【摘要】：由于无线传感器网络规模大、节点硬件能力受限、工作环境复杂等多方面限制,使得WSN网络传输安全受到挑战。本文对WSN密钥分配算法和路由安全协议进行了深入的研究,旨在设计出安全有效的密钥分配算法和路由安全协议。本文介绍了WSN网络体系的结构和特点,以此为基础针对WSN进行了安全需求和现状的分析,着重总结了常见的安全攻击以及相应的解决手段。其次,特别针对几种经典WSN路由协议进行了性能上的分析和安全性上的总结。通过总结和综合经典密钥分配算法,本文设计出了一种基于信用评价机制的混合型密钥管理方案。文中将信用评价机制引入密钥分配的过程中,之后将所有节点分为黑名单节点、安全节点和不安全节点三种类型。并且通过对于基于网格的密钥分配算法和地理位置算法不同的混合方式提出性能不同的模式一和模式二。对安全节点使用连通性高抗而俘获性低的模式一,同时,对不安全节点使用连通性抗低而俘获性高的模式二。对于不同安全性的节点,给予不同的密钥建立模型,从而实现了密钥分配算法性能的动态性。文中,从初始密钥分配过程、直接密钥建立过程和间接密钥建立过程对该算法进行了阐述。通过仿真分析可知:模式一全局连通性为0.2,密钥建立平均路径长度小于1.48,被抗俘获的概率在被俘获节点比例超过0.9后开始迅速增加。模式二全局连通性为0.02,密钥建立平均路径长度大于2.6,被抗俘获的概率在被俘获节点比例在0.2和0.9左右后开始迅速增加。方案使得密钥分配算法连通性、密钥建立路径长度以及抗俘获性都具有动态性。使得整个网络在比较安全的状态下连通性好,可以高效进行通信;而在整个网络安全性较低的情况下,网络整体的抗俘获性增强,从而防止密钥泄漏。GAF协议通常遭受虚假路由、选择性转发和女巫攻击三种类攻击。为了防止虚假路由攻击,本文提出将“基于信用评价机制的混合型密钥管理方案”运用到GAF协议,并对路由消息进行加密和认证。针对选择性转发攻击,本文设计了安全簇头选举算法。针对女巫攻击,本文算法使用哈希函数单向性来实现消息认证,从而防止恶意节点伪装。改进后的SGAF协议在一定程度上解决了GAF协议所面对的三大攻击问题。通过仿真分析可知,SGAF协议增加的能耗为GAF的8.3%,在可以接受的范围之内。
[Abstract]:Due to the large scale of wireless sensor network, limited hardware capability of nodes and complex working environment, the security of WSN network is challenged. In this paper, the WSN key distribution algorithm and routing security protocol are studied in order to design a secure and effective key distribution algorithm and routing security protocol. This paper introduces the structure and characteristics of WSN network architecture, analyzes the security requirements and current situation of WSN, and summarizes the common security attacks and corresponding solutions. Secondly, the performance and security of several classical WSN routing protocols are analyzed and summarized. By summing up and synthesizing the classical key distribution algorithm, a hybrid key management scheme based on credit evaluation mechanism is designed in this paper. In this paper, credit evaluation mechanism is introduced into the process of key distribution, and then all nodes are divided into three types: blacklist node, secure node and insecure node. In addition, we propose different performance modes one and two by combining the grid based key distribution algorithm and geographical location algorithm. A model with high connectivity and low trapping is used for secure nodes, while mode two is used for insecure nodes with low connectivity and high capture. For the nodes with different security, different key models are given to realize the dynamic performance of the key distribution algorithm. In this paper, the algorithm is described from the initial key distribution process, the direct key establishment process and the indirect key establishment process. The simulation results show that the global connectivity of mode is 0.2, the average path length of key establishment is less than 1.48, and the probability of resisting capture begins to increase rapidly when the proportion of captured nodes exceeds 0.9. The global connectivity of mode 2 is 0.02, the average path length of key establishment is more than 2.6, and the probability of resisting capture begins to increase rapidly after the proportion of captured nodes is about 0.2 and 0.9. The scheme makes key distribution algorithm connectivity, key establishment path length and anti-capture dynamic. In the case of low security, the anti-trapping ability of the whole network is enhanced, so as to prevent the key leakage. GAF protocol is usually subject to false routing. Selective forwarding and witch attacks are three types of attacks. In order to prevent false routing attacks, this paper proposes a hybrid key management scheme based on credit evaluation mechanism, which is applied to GAF protocol, and encrypts and authenticates routing messages. For the selective forwarding attack, a secure cluster head election algorithm is designed in this paper. In order to prevent malicious nodes from camouflage, this algorithm uses hash function one-way to implement message authentication. To some extent, the improved SGAF protocol solves the three major attack problems faced by the GAF protocol. The simulation results show that the increased energy consumption of SGAF protocol is 8.3% of GAF, which is within the acceptable range.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN918.4;TP212.9

【参考文献】