基于定性评估的WSN节点捕获攻击被动检测方法

Abstract

由于通常部署于外界，WSN节点易于被敌手捕获。传统捕获攻击的监测方法主要有基于缺席时间的监测及被动入侵检测两类，前者需要额外的通信开销，而后者则需要对网络整体信号强度进行统计分析，对单个节点入侵的识别通常不够敏感。本文利用云模型定性知识与定量数值之间的不确定性转换能力，对WSN节点之间通信中信号强度进行实时统计，建立信号强度云模型，得出节点是否遭遇入侵的定性判断，进而对可疑节点一段时间内信号强度进行分析，判断是否遭遇捕获攻击。仿真实验证明，该方法能够较大程度地提高检测的准确度，且误报率较低。 Since the nodes of WSNs are always deployed on the outside, nodes are easy to be captured. The traditional detection approaches of capture attack can be categorized as approaches based on time of absence and approaches based on passive intrusion detection. The former requires extra communication cost, and the latter needs to carry on the statistical analysis of the whole network signal strength. In this paper, the qualitative and quantitative uncertainty conversion ability of cloud model is used to evaluate the signal strengths among WSN nodes real-time. Normal cloud models are built based on the evaluation. The qualitative judgments of nodes are made, and the capture attacks in WSNs can be detected in time. Simulation results show that, this method can greatly improve the detection accuracy, and that the false alarm rate is low.