A Koopman-DMPC Strategy for Nonlinear Cyberphysical Systems Under Bounded Communication Delays

As technology advances, cyber and physical systems have become increasingly integrated, making the control of these integrated systems a critical challenge. This work investigates the consensus control problem for nonlinear cyberphysical systems (CPSs) subject to bounded communication delays. In particular, a novel consensus control strategy is proposed by integrating the Koopman operator theory with distributed model predictive control (DMPC). In this framework, each agent in the physical layer uploads its observed data to a cloud platform, where a data-driven linear model tailored for model-based control is constructed using the Koopman operator. Subsequently, optimal consensus control performance is achieved by minimizing the deviation between the DMPC input and a predefined optimal consensus input. Furthermore, under appropriate conditions, the recursive feasibility and the consensus analysis of the proposed control strategy are rigorously established. Finally, to validate the effectiveness of the method, experiments are conducted on three nonlinear systems with varying structures. The results demonstrate that the proposed strategy achieves superior consensus performance and robustness compared to traditional DMPC.