无线传感器网络KIPSO欺骗攻击检测模型

摘要

针对无线传感器网络通常处于开放性环境中，极易遭受外界欺骗攻击这一问题，设计了一种基于改进粒子群优化算法的K均值欺骗攻击检测模型。该模型将欺骗攻击检测描述为一个统计假设检验，基于接收信号强度空间与物理位置的相关性，利用不同位置节点接收信号强度的差异进行攻击检测；对接收信号强度进行聚类分析，计算类中心之间的距离，通过阈值检测判断节点是否受到欺骗攻击。仿真结果表明，KIPSO欺骗攻击检测模型能在提高检出率、增强报警可信度的同时，有效解决传统聚类算法陷入局部极值的问题。%As wireless sensor networks(WSNs)are usually deployed in open environment,they are vulnerable to the spoofing attack. A K-means based on Improved Particle Swarm Optimization(KIPSO)spoofing attack detection mod⁃el is designed to solve this problem. The model described the spoofing attack detection as a statistical hypothesis testing. Take advantage of the received signal strength(RSS)differences between locations to detect attack,which based on the spatial correlation of the location and the RSS space. Using KIPSO algorithm cluster the RSS in attack detection phase. Then calculate the distance between two centroids. Eventually,judging whether spoofing attack ex⁃ist or not via the trained threshold. Simulation results indicate that KIPSO spoofing attack detection model can not only improve detection rate or strengthen the alarm credibility,meanwhile,effectively solve problem of traditional clustering algorithm trapped in local minima.