Distributed model predictive control of virtual coupled trains based on NDOB & APF

Virtual coupling is a new train operation mode, in which trains form a logical whole based on vehicle-to-vehicle communication, which further shortens the traveling interval and improves transportation efficiency. The system uncertainty and external disturbances in the actual line environment have a large impact on the performance of train operation control, therefore, the safe and efficient tracking control problem of virtual coupled trains considering uncertainty disturbances is investigated. First, a nonlinear dynamic model of trains incorporating uncertain disturbances was constructed. Subsequently, a distributed model predictive control (DMPC) method based on nonlinear disturbance observer (NDOB) and artificial potential field (APF) was designed. Finally, a three-train section-tracking scenario simulation experiment was devised using real train and line data. Compared with other control methods, simulation results validated the effectiveness and superiority of the proposed control approach. It achieves stable and efficient control while ensuring safety, demonstrating higher control accuracy and stronger disturbance rejection capabilities.