隐私保护下的多智能体系统弹性一致性控制
Resilient consensus control of multi-agent systems under privacy protection
张冲 1伍益明 1徐明 1郑宁1
作者信息
- 1. 杭州电子科技大学网络空间安全学院,杭州 310018
- 折叠
摘要
针对多智能体系统通信网络遭受欺骗攻击与隐私窃取问题,本文提出了一种新的具备隐私保护能力的多智能体系统弹性一致性算法.首先考虑到多智能体系统在信息传输环节直接将真实状态信息暴露给外界的问题,设计了一种基于状态分配思想的节点状态信息处理机制,来保证系统中节点状态信息隐私的安全性.其次考虑到外部攻击者对系统收敛的影响,在状态分配机制上进一步引入了检测环节,从而保证系统安全收敛.然后通过数学理论分析证明了算法能够有效保护节点初始状态信息隐私的同时能够保证系统实现弹性一致性收敛.最后通过数值仿真实验与对比实验,进一步验证了本文算法的有效性.
Abstract
This paper presents a new privacy-preserving resilient consensus algorithm for multi-agent systems whose communication network suffers from deception attacks and privacy theft.First,considering the multi-agent system directly exposes the real state information to the outside in the information transmission process,a node state information processing mechanism based on state allocation is designed to ensure the security of node state information privacy in the system.Secondly,considering the influence of external attackers on system consensus,the detection step is further introduced into the state allocation mechanism to ensure the safe consensus of the system.Third,mathematical theory analysis proved that the algorithm can effectively protect the privacy of node initial state information and ensure the system achieves resilient consensus.Finally,the effectiveness of this algorithm is further verified by numerical simulation and comparison experiments.
关键词
多智能体系统/弹性一致性/欺骗攻击/隐私保护/网络安全Key words
multi-agent systems/resilient consensus/deception attack/privacy-preserving/cyber security引用本文复制引用
基金项目
国家自然科学基金(62073109)
浙江省公益技术应用研究项目(LGF21F020011)
出版年
2024
NETL
NSTL
万方数据