Zhang Dong-Yu, Li Xiao-Jian
IEEE Trans Cybern. 2024 Dec;54(12):7404-7415. doi: 10.1109/TCYB.2024.3467272. Epub 2024 Nov 27.
This article studies the secure consensus control problem of multiagent systems (MASs) with a nonzero input leader subject to denial-of-service (DoS) attacks. The introduction of backup topologies makes it possible for MASs to actively defend against DoS attacks. Subsequently, a novel active defense method consisting of two types of state observers, an adaptive topology switching mechanism, and switching controllers is proposed, which can ensure the leader-follower bound consensus even if DoS attacks hinder the interaction between agents. Within such a defense framework, the switching mechanism, driven by the predefined performance index and designed monitoring function, can automatically search for a healthy communication graph among backup topologies. Concurrently, the observer-based switching control strategy will be modified to match the corresponding topology, in which the universal observer and controller parameters in different topologies are obtained by solving linear matrix inequalities. It should be highlighted that the developed defense scheme not only removes the limitations of existing results on the duration and frequency of DoS attacks but also ensures the same upper bound of consensus error before and after DoS attacks. Finally, several simulation examples for different systems illustrate the efficiency and superiority of the theoretical results.
本文研究了具有非零输入领导者的多智能体系统(MASs)在遭受拒绝服务(DoS)攻击时的安全一致性控制问题。备用拓扑结构的引入使MASs能够主动抵御DoS攻击。随后,提出了一种由两种状态观测器、一种自适应拓扑切换机制和切换控制器组成的新型主动防御方法,即使DoS攻击阻碍了智能体之间的交互,该方法也能确保领导者-跟随者有界一致性。在这样的防御框架内,由预定义性能指标和设计的监测函数驱动的切换机制可以在备用拓扑结构中自动搜索健康的通信图。同时,基于观测器的切换控制策略将被修改以匹配相应的拓扑结构,其中不同拓扑结构中的通用观测器和控制器参数通过求解线性矩阵不等式获得。需要强调的是,所提出的防御方案不仅消除了现有结果对DoS攻击持续时间和频率的限制,而且确保了DoS攻击前后一致性误差的相同上界。最后,针对不同系统的几个仿真例子说明了理论结果的有效性和优越性。