Event-Triggered Quantized Consensus of Linear Multiagent Systems Under Asynchronous Denial-of-Service Attacks

This article investigates the event-triggered consensus problem for a general linear multiagent system (MAS) under limited feedback bit rates. Moreover, the interagent communication edges of the MAS may be blocked by multiple attackers in an asynchronous manner. Since each agent has to quantize the feedback information into finite-bit symbols before transmissions, an asynchronous attack may break the quantization synchronization among neighboring agents. To mitigate the effects of asynchronous attacks, we design a distributed dynamic encoding–decoding scheme, under which the quantization range and the number of occupied bits of each transmission of every communication edge are independently determined and may vary by DoS attacks. Furthermore, we propose an event-triggered quantized consensus protocol, under which distributed event-triggers determine information transmission over each communication edge. Under the proposed methods, we provide a sufficient average bit rate condition to ensure the asymptotic consensus of the concerned MAS. That bit rate condition depends on the dynamics of agents, the communication topology and denial-of-service (DoS) attacks. The proposed methods still works in MASs with bounded network delay and asynchronous DoS attacks. A numerical example is provided to verify the effectiveness of obtained results.

I. Introduction

Consensus control is a critical task of multiagent systems (MASs) [1], [2], [3]. It requires all agents to reach a common value based on a consensus protocol. Research on the consensus problem of MASs has been well studied [4], [5], [6]. The obtained results are usually built on the common assumption of ideal communication, i.e., agents can exchange information continuously, perfectly and in an error-free way. Unfortunately, this assumption may be broken in reality. Since interagent communication often suffers from the limited bandwidth constraint and malicious attacks, we focuses on the quantized consensus for general linear MASs under denial-of-service (DoS) attacks.