Data-Driven Safety-Critical Cooperative Control For Multiple-unit Trains

Cooperative control is a fundamental problem in the automatic train operation of Multiple-unit (MU) trains. Many recent studies concentrate on multi-point longitudinal dynamics of trains to achieve asymptotic stability of cooperation. However, the latent influence on coupler safety has not been fully addressed. An unpleasant transient response of the cooperation process may accelerate coupler aging and finally lead to coupler fracture and operational safety incidents. In this paper, we designed a data-driven safety-critical control strategy for MU trains to achieve both cooperation and coupler safety. Specifically, a robust adaptive control barrier function (RaCBF) is constructed for the coupler safety of MU trains, guaranteeing the coupler length within a reasonable range for the whole process of cooperation. Moreover, set membership identification (SMI) technique is combined with the adaptive law to tackle the system parameter uncertainties and increase the flexibility of the controller. The proposed RaCBF is synthesized with a traditional feedback cooperative controller via quadratic programming (QP) to achieve coupler safety and cooperative tracking performance. Finally, simulations compared with the traditional controller demonstrate the effectiveness of the proposed controller.