计算机科学
计算机安全
信息隐私
控制(管理)
隐私保护
控制系统
密码学
计算机网络
方案(数学)
电子邮件
数据完整性
稳健性(进化)
弹性(材料科学)
服务器
作者
Zheng Zhang,Wenyi Wang,Shiming Chen,Xiujuan Zhao,Yuanshi Zheng
标识
DOI:10.1109/tase.2026.3745819
摘要
This paper is the first to investigate whole-process privacy preservation and resilient consensus control for sampled-data first/second-order heterogeneous multi-agent systems under the coexistence of dynamic malicious attacks and eavesdroppers. To address the persistent state contamination induced by alternating attacked and recovered nodes in dynamic attacks, a distributed HDP-MSR filtering algorithm is proposed. By introducing virtual position variables for second-order agents, an extended state model is constructed to unify the dynamic description of heterogeneous MASs. Under sampling period and control gain constraints, the system update is proved to satisfy the convex combination property, and a sufficient topological condition is derived to guarantee resilient consensus. Furthermore, a whole-process privacy preservation mechanism integrating public key encryption and differential information interaction is established. It achieves whole-cycle privacy protection for agents’ initial and dynamic states without degrading consensus accuracy. Theoretical analysis shows that normal agent states can always remain within a safe range and reach asymptotic consensus, effectively resisting both dynamic malicious attacks and eavesdropping. Simulation results validate the superiority of the proposed algorithm in attack resilience and whole-process privacy preservation.
科研通智能强力驱动
Strongly Powered by AbleSci AI