电子学报2026，Vol.54Issue(3)：981-999,19.DOI:10.12263/DZXB.20251088
拒绝服务攻击下高阶链式结构多智能体系统的有限时间安全控制
Finite-Time Secure Control of High-OrderChain-Structured Multi-Agent Systems under Denial-of-Service Attacks
摘要
Abstract
This paper focuses on the finite-time secure control problem of high-order chain-structured multi-agent sys-tems under denial-of-service(DoS)attacks,and conducts systematic modeling,algorithm design and theoretical verification.Firstly,a high-order chain dynamic model of the leader-follower framework is established,which accurately describes the point-to-point DoS attack scenario and defines the core constraint that attacked agents cannot send or receive information,thus restoring the characteristics of communication interruption under real cyber attacks.To address the issues of communication link failure and state information loss caused by attacks,a distributed finite-time observer is designed to achieve fast and accu-rate estimation of the leader's state and input information,effectively compensating for the information loss due to communica-tion interruption.Meanwhile,the impact of DoS attacks on the convergence time of the observer is quantitatively analyzed,and it is rigorously proved that the observer can maintain the expected convergence performance as long as the attack period is upper bounded.Based on the backstepping design procedure and finite-time control theory,a secure control protocol adapted to the characteristics of high-order chain systems is proposed,and the system stability is verified via Lyapunov stability theory,ensuring that the system can achieve finite-time stability despite attack disturbances.To validate the effectiveness of the pro-posed method,numerical simulations are carried out in a scenario with a total DoS attack duration of 3 seconds,using a wheeled mobile robot formation as the simulation object and constructing a high-order chain multi-agent system with one lead-er and five followers.The results show that the designed observer can quickly restore accurate estimation of the leader's infor-mation within 0.1 seconds during attack intervals,the state observation error converges to zero in about 6 seconds,and the tracking error achieves complete convergence in about 7 seconds,significantly improving the system's attack-resistant recov-ery speed.Compared with the existing observer-based distributed finite-time control methods,the proposed scheme exhibits ob-vious advantages in convergence speed,anti-interference ability and tracking accuracy.Furthermore,noise robustness and pa-rameter sensitivity tests verify that the system can operate stably under measurement noise interference,and small fluctuations in control parameters have little effect on convergence performance,indicating good engineering applicability.The research re-sults improve the theoretical system of secure control for high-order chain multi-agent systems under DoS attacks,and can pro-vide technical support for finite-time secure cooperative control of key equipment and systems vulnerable to cyber attacks,such as unmanned aerial vehicle formations,intelligent robot swarms,and smart grids,enhancing the operational reliability and survivability of systems in complex network environments.关键词
多智能体系统/DoS攻击/有限时间控制/分布式观测器/高阶链式/安全控制Key words
multi-agent systems/DoS attacks/finite-time control/distribute observer/high-order chained/secure control分类
信息技术与安全科学引用本文复制引用
王靖瑶,宋健,郭景华,曾泽钦..拒绝服务攻击下高阶链式结构多智能体系统的有限时间安全控制[J].电子学报,2026,54(3):981-999,19.基金项目
厦门市自然科学基金(No.3502Z202471004) （No.3502Z202471004）
国家自然科学基金(No.62473323,No.52372419) Xiamen Natural Science Foundation(No.3502Z202471004) （No.62473323,No.52372419）
National Natural Science Founda-tion of China(No.62473323,No.52372419) （No.62473323,No.52372419）
