Abstract This paper investigates the sampled-data-based consensus problem of multi-agent systems (MASs) under asynchronous denial-of-service (DoS) attacks. In order to describe asynchronous DoS attacks, a new definition of complete DoS attack and novel double-layer switched systems are proposed. A complete DoS attack refers to a DoS attack that consists of several consecutive successful DoS attacks. While a successful DoS attack denotes an attack that can break the connected communication topology into several isolated subgraphs. Based on this, the original system is transformed into a double-layer switched systems with a stable mode and several unstable modes. It should be pointed out that each unstable subsystem is also composed of finite second-level unstable subsystems that represent consecutive successful DoS attacks. Moreover, a new double-mode-dependent Lyapunov function (DMDLF) method is employed to obtain the lower and upper bounds of the corresponding average dwell time (ADT) of subsystems. It is proved that the consensus of MASs under asynchronous DoS attacks can be achieved by using the feedback consensus controllers which can be designed simultaneously. Finally, an illustrative example is provided to illustrate the effectiveness of the results proposed in this paper.

中文翻译：

---

异步拒绝服务攻击下基于采样数据的多代理系统共识

摘要 本文研究了异步拒绝服务 (DoS) 攻击下多代理系统 (MAS) 的基于采样数据的共识问题。为了描述异步DoS攻击，提出了完全DoS攻击和新型双层交换系统的新定义。一次完整的 DoS 攻击是指由多次连续成功的 DoS 攻击组成的一次 DoS 攻击。虽然成功的 DoS 攻击表示可以将连接的通信拓扑分解为几个孤立的子图的攻击。在此基础上，将原系统转化为具有一个稳定模式和多个不稳定模式的双层切换系统。需要指出的是，每个不稳定子系统也由代表连续成功的DoS攻击的有限二级不稳定子系统组成。而且，采用一种新的双模依赖李雅普诺夫函数 (DMDLF) 方法来获得子系统相应平均驻留时间 (ADT) 的下限和上限。证明了在异步 DoS 攻击下 MAS 的共识可以通过使用可同时设计的反馈共识控制器来实现。最后，提供了一个说明性的例子来说明本文提出的结果的有效性。