Event-triggered model predictive control with high-gain disturbance observers for cooperative control of virtual coupling trains in metros

To enhance the tracking accuracy and control stability of virtually coupled train sets (VCTS) in metros under uncertain disturbances, such as track vibrations and adverse weather, a distributed event-triggered model predictive control (MPC) strategy based on a high-gain disturbance observer (HGDO) is developed, with consideration of both the computational efficiency and disturbance rejection capability of the control system. First, a cooperative control cost function is designed based on the dynamic model of virtual coupling trains. To alleviate the high computational burden of conventional MPC, an event-triggered mechanism is introduced, where adaptive trigger threshold sets are designed according to train's real-time speed, thereby reducing the computational load. Simultaneously, an HGDO is employed to observe and compensate external disturbances in real time, enhancing both disturbance rejection and control precision. The effectiveness and superiority of the proposed approach are validated through cooperative train control simulations conducted in MATLAB/Simulink. The results demonstrate that the proposed approach achieves computational efficiency improvements of 54.23%, 41.79%, and 29.35% under three distinct trigger threshold sets, compared to conventional MPC. Furthermore, relative to the advanced Input Blocking Robust MPC, the proposed method reduces the speed root mean square error by 26.43%, the distance root mean square error by 27.14%, and the average solution time by 18.75%, demonstrating its strong potential for practical metro applications.