基于CBMeMBer滤波器的多攻击检测方法

Multi-attack Detection Approach Based on CBMeMBer Filter

对具有大量节点的信息物理系统(CPS)而言,多攻击的同时在线检测是一个颇具挑战性的问题。一方面,多个不同攻击者可采用不同的攻击策略与发动不同类型的攻击以规避检测;另一方面,CPS不仅需检测攻击的数目,还需同时检测出被每个攻击者攻击的节点。为此,提出了一种基于随机有限集的多攻击检测方法。通过采用多伯努利随机集对攻击行为和CPS检测机制进行建模,并使用势均衡多目标多伯努利滤波器对模型滤波,所提方法可实现对CPS中多攻击的同时在线检测。

Simultaneous online detection of multiple attacks is one of challenging problems for cyber physical system(CPS)with numerous nodes.On the one hand,multiple different attackers can escape from detection by adopting different attack strategies and launching different types of attacks.On the other hand,CPS detects not only the number of attacks,but also each node attacked by each attacker simultaneously.To solve this problem,based on random finite set(RFS)theory,a novel approach for multi-attack detection is proposed.By utilizing the balanced multi-target multi-Bernoulli set to model the behavior of attacks and CPS detection schemes,and using the cardinality balanced multi-target multi-Bernoulli filter to filter the formulated model,the proposed approach achieves the simultaneous detection of multiple attacks in a CPS.