文献类型：期刊文献

英文题名：Secure Fully Distributed Event-Triggered Consensus of Multi-Agent Systems Against Distributed Sequential Scaling Attacks

作者：He, Wangli[1];Li, Shifen[1];Ge, Xiaohua[2];Qian, Feng[1]

机构：[1]East China Univ Sci & Technol, Key Lab Smart Mfg Energy Chem Proc, Minist Educ, Shanghai 200237, Peoples R China;[2]Swinburne Univ Technol, Sch Sci Comp & Engn Technol, Melbourne, Vic 3122, Australia

年份：2024

卷号：20

期号：2

起止页码：1764

外文期刊名：IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS

收录：;EI(收录号：20232314199405);WOS:【SCI-EXPANDED(收录号:WOS:001129361500006)】；

基金：This work is supported by National Natural Science Foundation of China (Basic Science Center Program: 61988101), Shanghai Pilot Program for Basic Research (22TQ1400100-3), Shanghai International Science & Technology Cooperation Program (21550712400), the Joint Fund of Ministry of Education for Equipment Pre-research (8091B022234).

语种：英文

外文关键词：Multi-agent systems; fully distributed consensus protocol; secure consensus; distributed sequential scaling attacks; fully distributed event-triggered mechanism

摘要：This article is concerned with the secure fully distributed event-triggered consensus problem of general linear multi-agent systems subject to distributed sequential scaling (DSS) attacks. First, a generic DSS attack model is proposed, which enables different attack strategies in terms of scaling factors, attack frequency and duration to be incorporated in various communication channels. Different from existing attack models, the DSS attack is a kind of scaling attacks in which attack sequences are characterized by the sequential attacks and with distributed attack strategies for different attack objectives. To resist the adverse effects of such DSS attacks and further reduce the unnecessary communication consumption of each communication channel, a channel-based dynamic event-triggered mechanism is next presented. Moreover, a fully distributed event-triggered consensus control protocol is developed such that the dependence of any global information of the network topology can be eliminated. Formal analysis criteria on the asymptotic convergence of the resultant consensus errors and Zeno-freeness are then derived, where the relationship of the triggering parameters, distributed scaling factors and attack constraints is explicitly expressed. Furthermore, an offline algorithm without any global information is provided to determine both the adaptive consensus protocol gain matrices and the triggering parameters. Therefore, the parameters calculated by this algorithm are applicable to multi-agent systems of different scales, which also confirms the flexibility and scalability of the proposed fully distributed event-triggered consensus control protocol. Finally, two simulation examples involving different scales of intelligent vehicles are given to validate the efficacy of the obtained theoretical results.