Output Synchronization of Unknown Linear Multiagent Systems: A Data-Driven Approach
Qiyu Wang, Juntao Zhang, Mingxia Gu, Abdujelil Abdurahman · 2025
This paper addresses the output synchronization problem for heterogeneous linear multi-agent systems (MASs) with unknown dynamics and external disturbances. Traditional model-based methods face limitations due to the unavailability of accurate system models, particularly under noise-corrupted environments. To overcome these challenges, a data-driven framework is proposed, eliminating the need for prior knowledge of the leader’s dynamics. By constructing an auxiliary system, the leader’s state matrix is estimated directly from sampled data, while data-based regulator equations are derived to resolve agent heterogeneity. The study establishes necessary and sufficient conditions for dataset informativity, ensuring the solvability of the output synchronization problem under noisy measurements. A distributed control protocol is designed, integrating observer-based leader state estimation and data-driven feedback gains to achieve consensus. Numerical simulations involving six heterogeneous followers validate the approach, demonstrating rapid convergence of follower outputs to the leader’s trajectory within finite time. The results highlight the method’s robustness to unknown dynamics, noise, and heterogeneous agent structures, thereby advancing practical data-driven strategies for MAS synchronization in scenarios with limited model knowledge and abundant data availability.