Zhang Dandan, Tang Yang, Ding Zhengtao, Qian Feng
IEEE Trans Cybern. 2021 May;51(5):2490-2503. doi: 10.1109/TCYB.2019.2910614. Epub 2021 Apr 15.
This paper focuses on the time-varying formation tracking issue for nonlinear multiagent systems (MASs). Based on the explicit characterizations of frequency, duration, and magnitude properties for deception attacks, a hybrid framework is proposed for time-varying formation tracking of nonlinear MASs. To realize the desired formation tracking performance under deception attacks, the distributed edge-based event-triggered communication strategies are proposed with Zeno-freeness. The designed strategies are resilient to deception attacks under some appropriate assumptions, to realize a predefined formation and simultaneously track the convex combination of leaders' states. The designed control strategies render that we do not need to detect when the deception attack happens. Furthermore, the obtained results can be deduced to deal with consensus/synchronization problems, target enclosing problems for MASs with one/multiple leaders, where the communication is attacked by malicious attackers. An example of time-varying formation tracking of unmanned aerial vehicles is provided to show the effectiveness of the obtained results.
本文聚焦于非线性多智能体系统（MASs）的时变编队跟踪问题。基于欺骗攻击的频率、持续时间和幅度特性的明确表征，提出了一种用于非线性MASs时变编队跟踪的混合框架。为了在欺骗攻击下实现期望的编队跟踪性能，提出了具有无芝诺特性的基于分布式边缘的事件触发通信策略。在一些适当假设下，所设计的策略对欺骗攻击具有弹性，以实现预定义的编队并同时跟踪领导者状态的凸组合。所设计的控制策略使得我们无需检测欺骗攻击何时发生。此外，所获得的结果可以推广到处理共识/同步问题、具有一个/多个领导者的MASs的目标包围问题，其中通信受到恶意攻击者的攻击。提供了一个无人机时变编队跟踪的例子来展示所获结果的有效性。
