欺骗攻击下多智能体网络的均方有界聚类一致性  

Mean-Square Bounded Cluster Consensus of Multi-Agent Networks Under Deception Attacks

作　　者：王柳 胡爱花[1] 江正仙[1] WANG Liu;HU Aihua;JIANG Zhengxian(School of Science,Jiangnan University,Wuxi 214122)

机构地区：[1]江南大学理学院,无锡易214122

出　　处：《系统科学与数学》2024年第3期634-647,共14页Journal of Systems Science and Mathematical Sciences

基　　金：江苏省自然科学基金(BK20181342);国家自然科学基金(61807016)资助课题。

摘　　要：针对非线性多智能体网络,研究了欺骗攻击下网络系统均方有界聚类一致性问题.首先,将所有网络节点分为不同子群.由于欺骗攻击的存在,控制信号可能会被错误信号代替,通过引入伯努利随机变量,衡量欺骗攻击是否成功.其次,设计一种采用牵引策略的分布式脉冲控制器,确保在存在欺骗攻击的情况下实现均方有界聚类一致性.进一步,利用图论、线性矩阵不等式和李雅普诺夫函数方法,给出了欺骗攻击下实现多智能体网络均方有界聚类一致性的充分条件.最后,通过仿真例子验证理论结果的可行性和有效性.This paper studies the mean-square bounded cluster consensus for non-linear multi-agent networks under deception attacks.Firstly,all network nodes are divided into different clusters.Considering that the control signal may be replaced by an error signal after being subjected to deception attacks,a Bernoulli random variable is introduced to represent the success of the deception attack.Secondly,a distributed impulsive controller employing pinning strategy is designed to ensure mean-square bounded cluster consensus in the presence of deception attacks.Furthermore,using the graph theory,linear matrix inequality and Lyapunov function method,sufficient conditions for realizing the mean-square bounded cluster consensus of multi-agent networks under deception attacks are given.Finally,a simulation example is supplied to verify the feasibility and effectiveness of the theoretical results.

关 键 词：多智能体网络 欺骗攻击 聚类一致性 牵引策略 脉冲控制 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]