This invention discloses a
levitation control method, equipment, and medium for conventional
maglev trains, belonging to the field of
maglev transportation control technology. The method constructs a control framework based on a dynamic model of the
levitation system and introduces an online residual
estimation and compensation mechanism to compensate for track irregularities, aerodynamic disturbances, and unmodeled dynamics under complex operating conditions. A lightweight data-driven model obtained through cloud-based large-
scale model knowledge
distillation is deployed in the onboard
edge computing unit to model the temporal characteristics of
system state, track features, and external environmental information, estimating disturbance residuals in real time and generating compensating control quantities that are superimposed on the control input. Simultaneously, the cloud continuously optimizes
model parameters based on historical operating data and issues updates, achieving iterative optimization of the control strategy. This invention can significantly improve the stability, dynamic response performance, and
operational safety of the
maglev train levitation system under high-speed operating conditions.