面向合谋攻击的多智能体系统弹性一致性控制
Resilient Consensus Control for Multi-agent Systems Against Collusive Attacks
-
摘要: 随着攻击的智能化发展, 攻击倾向于合谋以提高隐蔽性并造成更大的破坏, 给多智能体系统在复杂环境中执行协同控制任务带来巨大挑战. 在攻击合谋能力受限的约束下, 设计抗合谋攻击的分布式隔离算法, 分析攻击零漏报的充分条件. 在此基础上, 提出基于攻击隔离的弹性控制算法, 确保正常智能体完成既定一致性任务, 提高多智能体系统的弹性和抗毁性.Abstract: With the development of intelligent attacks, attacks trend to collude to improve the stealthiness and to cause more damages, which brings great challenges for multi -agent systems to execute coordinated control tasks under complex environment. Under the constrained ability of attack collusion, the distributed isolation algorithm against collusive attacks is designed, and a sufficient condition is analyzed to accomplish zero missed attack alarm. Base on this, the resilient control algorithm based on attack isolation is proposed to assure the accomplishment of assigned consensus task by normal agents and to improve the resilience and invulnerability of multi-agent systems.
下载: