Intelligent Consensus Asymptotic Tracking Control for Nonlinear Multiagent Systems Under Denial-of-Service Attacks

This work considers the adaptive output feedback consensus asymptotic tracking control problem for full-state-constrained nonlinear multiagent systems (MASs) under the denial-of-service (DoS) attacks and the actuator faults. Since MASs are regardered as a network, where each agent is considered as a network node, the DoS attacks that cut network communication links between agents are studied in this paper. Moreover, to compensate the unknown actuator fault signal, the Nussbaum function method is introduced into the controller design process for each agent. Due to the researched nonlinear MASs contain unknown nonlinearities and the system states of each agent are unmeasurable, an estimator based on fuzzy logic systems (FLSs) is constructed for each agent. Further, the secure control strategy based on the barrier Lyapunov functions (BLFs) is proposed for the researched nonlinear MASs, which guarantees the consensus tracking errors asymptotically converge to zero, all closed-loop signals remain bounded and the full-state constraint requirements are not broken. In the illustrative example, the proposed secure control strategy is applied to the MASs composed of multiple forced damped pendulums (FDPs). Note to Practitioners— In the framework of network communication, how to eliminate the impact of cyber attacks on nonlinear MASs that rely on communication links to transmit information is critical. Furthermore, due to practical systems are inevitably subject to multiple physical and environmental constraints, the problem of dealing with system constraints can not be ignored. Therefore, a secure control algorithm based on the BLFs is proposed to eliminate the effects of the DoS attacks on the researched nonlinear MASs and ensure that the full-state constraints are not broken. To realize the output feedback control and approximate the unknown nonlinearities of the studied system, the estimators combining with the intelligent approximation techniques based on FLSs are developed. The Nussbaum function is introduced into controller to solve the actuator faults problems for each agent.

I. Introduction

The burgeoning application potential of MASs has prompted a surge in interest regarding the consensus control problem for MASs [1], [2], [3], [4]. Given that numerous real-world systems exhibit intricate nonlinear dynamics, the focus of research has recently shifted from linear MASs to nonlinear MASs [5], [6], [7], [8]. According to the different communication relationships between agents, consensus control can be divided into two categories: leaderless consensus control [9] and [10] and leader-following consensus control [11], [12], [13]. It is shown in [1] and [3] that the leader-following consensus control is more practical than the leaderless consensus control. Due to the presence of the communication topology and the limited sensing capability of sensors, all agents under the leader-following control framework should be completely distributed, which means each agent does not require the information of all agents, relies only on the partial information from its neighbors. It is worth noting that when the neighbor’s information is maliciously attacked, the consensus performance of nonlinear MASs will be threatened. Therefore, how do we maintain the ideal consensus performance of the attacked nonlinear MASs is crucial.