Resilient adaptive event-triggered containment control of nonlinear multi-agent system under concurrent DoS attacks and disturbances

Abstract

This paper presents a secure containment control problem of nonlinear Multi-Agent Systems (MASs) under aperiodic Denial of Service (DoS) attacks and external disturbances simultaneously. A novel adaptive neural network (NN)-based event-triggered control is considered that uses the nonlinear estimator to predict the state of other agents. Since data access is denied during DoS attacks, the overall system switches between two modes of stable and unstable containment behaviours. Therefore, the maximum of attack duration and frequency is determined such that the overall system evolution leads to containment convergence in the presence of DoS attacks. We proposed an adaptive NN-based distributed disturbance observer to estimate external disturbances in a nonlinear system's dynamics. The state estimator predicts neighbouring agents' states, and each agent's input and event times are determined without monitoring other agents. The directed graph topology is used to determine data exchange among agents instead of an undirected graph that reduces implementation conditions. Zeno-free behaviour is also proved by analysis of the system. Eventually, the numerical simulation of the proposed approach is shown.

Abbreviations: DoS attacks, Containment control of multi-agent system

Keywords:

• Containment control
• denial of service
• disturbances
• event-triggered control
• multi-agent system
• neural network
• nonlinear dynamics

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data sharing is not applicable to this article as no new data were created or analysed in this study.

Additional information

Funding

This work is based upon research funded by Iran National Science Foundation (INSF) under project No. 4015920.