As the input of the aggregate function is vulnerable to attack, an attack detection algorithm based on F-distribution is proposed. Based on the distributed model of data aggregation, the variable obeyed to F-distribution was constructed. According to the constant false alarm rate and the statistical properties of the variable, the threshold value of the algorithm can be obtained and the requirements for priori information are not so important. Simulation results show that the attack detection performance of the new algorithm is better, and its ability against noise jamming is more powerful than other algorithms.%针对汇聚函数的输入值易受敌方攻击的问题,提出一种基于F分布的无线传感器网络攻击检测方法.该方法在分布式数据汇聚模型的基础上构造服从F分布的统计量,在虚警率恒定的情况下,可根据统计量的统计特性确定攻击检测的门限阈值,放宽了对节点感知数据先验信息的要求.计算机仿真结果表明新方法的攻击检测率较高,且对噪声干扰的稳健性较强.
展开▼
