Hierarchical Distributed Data‐Driven Resilient Control of Hybrid‐Order MASs With Communication Link Faults

ABSTRACT The leader‐following consensus problem is studied in this article for heterogeneous hybrid‐order multi‐agent systems (MASs) subject to communication link faults. Unlike most existing studies focusing on MASs with identical‐order dynamics, this article considers the heterogeneous hybrid‐order MASs featuring both first‐ and second‐order nonlinear dynamics. A hierarchical distributed data‐driven resilient control strategy is proposed, featuring a two‐layer structure. Specifically, a resilient adaptive distributed observer is designed in the upper layer to estimate the leader's states despite communication link faults, while a data‐driven adaptive tracking controller with order‐dependent mechanisms is designed in the lower layer for followers with hybrid‐order dynamics. The proposed control framework not only relaxes the stringent topological requirements of the existing distributed methods by decoupling the state estimation from the tracking control but also eliminates the dependence on accurate system models through the designed data‐driven method. Stability analysis of the proposed resilient adaptive distributed observer demonstrates that the estimation errors converge to zero exponentially in the presence of communication link faults, based on which it is further proved that the tracking errors are uniformly ultimately bounded. Finally, numerical and comparative simulations are conducted to validate the effectiveness of the proposed control strategy.