基于信任估计的恶意节点检测策略

发布时间：2018-03-12 11:58

本文选题：无线传感器网络　切入点：信任模型　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：无线传感器网络的有限能量和广播特质使无线传感器网络更容易受到攻击。因此,设计适合无线传感器网络特性的安全机制,抵制各种攻击是确保网络传输可靠数据的重要问题。目前信任模型已引用到无线传感器网络的安全机制中。然而,信任模型的自身缺陷也可成为被攻击的对象。并且无线传感器网络中的恶意节点还会影响用户的判断。针对这些问题本文具体工作如下:首先,提出一种基于信任模型的无线传感器网络恶意节点检测策略。该策略通过引入指数衰减因子来刻画节点动态行为的变化,同时将交互失败信息的比重在信任评估中逐渐增大,使所建立的模型对策略性攻击具有较强的抵御能力。针对可能存在的恶意推荐问题,通过计算评估节点和推荐节点对公共交互节点的信任评估的平均差来建立推荐可信度,并将其作为推荐信任值的权重以降低恶意推荐的影响。另外,总体信任值评估时,将交互经验充分性因子作为动态权重因子来平衡直接和推荐信任值。实验结果验证了该策略对网络中恶意节点有较强的抵御能力和有较高的准确性。其次,设计了一种基于加权信任估计的恶意节点检测策略。该策略先运用双阈值进行事件检测,并通过调整双阈值来权衡事件检测准确度和事件误检率。同时,通过设计一种事件边界检测算法来降低对事件边界节点及其邻居的误判。最终提高恶意节点检测率和降低误判率。仿真实验验证了该策略的有效性。
[Abstract]:The limited energy and broadcast characteristics of wireless sensor networks make wireless sensor networks more vulnerable to attack. Resisting all kinds of attacks is an important problem to ensure the transmission of reliable data in the network. At present, the trust model has been referenced in the security mechanism of wireless sensor networks. However, The defects of the trust model can also be the target of attack, and the malicious nodes in the wireless sensor network will also affect the user's judgment. The specific work of this paper is as follows: first of all, This paper proposes a malicious node detection strategy based on trust model in wireless sensor networks, which describes the dynamic behavior of nodes by introducing exponential attenuation factor, and increases the proportion of mutual failure information in trust assessment. The model is able to resist the strategic attack. In view of the possible malicious recommendation problem, the recommendation credibility is established by calculating the average difference between the evaluation node and the recommendation node in the trust evaluation of the common interactive node. It is used as the weight of the recommended trust value to reduce the impact of malicious recommendation. In addition, when the overall trust value is evaluated, The interactive empirical adequacy factor is used as the dynamic weight factor to balance the direct and recommended trust values. The experimental results show that the strategy has strong resistance and accuracy to the malicious nodes in the network. Secondly, A malicious node detection strategy based on weighted trust estimation is designed, in which two thresholds are first used to detect events, and the accuracy of event detection and the false detection rate of events are measured by adjusting double thresholds. An event boundary detection algorithm is designed to reduce the misjudgment of event boundary nodes and their neighbors. Finally, the detection rate of malicious nodes is improved and the false judgment rate is reduced. The simulation results show the effectiveness of the strategy.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP212.9;TN915.08

【参考文献】