Data-Driven Learning for Resilient Synchronization and Parameter Estimation of Heterogeneous Nonlinear Multiagent Systems
Wang Yang,Jiuxiang Dong
DOI: https://doi.org/10.1109/tase.2023.3335238
IF: 6.636
2023-01-01
IEEE Transactions on Automation Science and Engineering
Abstract:This article is focused on the output synchronization or tracking problem for heterogeneous uncertain nonlinear multiagent systems (MASs) affected by denial-of-service (DoS) attacks, which realizes the two joint objectives, namely output synchronization and parameter estimation. Different from the recent works, where neighbors of the leader can access the state of the leader, only part output information of the leader is available for its neighbors in this article. In addition, the system matrix of the leader only is known to neighbors of the leader. In order to estimate the state and system matrix of the leader and resist the effect of DoS attacks, a distributed resilient observer is proposed by employing a jointly observable condition. Moreover, to derive precise adaptive learning of unknown parameters in uncertainties, a data-driven learning method is introduced. Consequently, a novel resilient output synchronization control framework is proposed, in which the interesting fact is that exponential convergence of both output synchronization errors and parameter estimation errors to zero can be arrived simultaneously. Finally, a simulation example is presented to reveal the effectiveness of the developed control scheme. Note to Practitioners —The multiagent output synchronization problem is studied in this paper, which can be applied to some many practical control systems with uncertain parameters. In many practical applications, owing to physical or geographical limitations, it is usually not easy for neighbors of the leader to directly measure the full state or output of the leader. A single agent might be able to obtain only partial output of the leader. In addition, the actual networked MASs are vulnerable to cyber attacks caused by malicious attackers, which can lead to output synchronization degradation or instability. Therefore, in this paper, an adaptive resilient synchronization control scheme is developed for MASs with parameter uncertainties under jointly observable condition. In simulations, a two-mass-spring system is used to illustrate the effectiveness of the proposed control scheme.