Cooperative MFAILC for Multiple Subway Trains With Actuator Faults and Actuator Saturation

Abstract

In this paper, a cooperative model free adaptive iterative learning control (CMFAILC) scheme is proposed for multiple subway trains subjected to actuator faults and actuator saturation. The controller design includes: the dynamic linearization technique is adopted to transform the subway train system into an equivalent dynamic linearization data model along the iteration axis, the CMFAILC controller is designed by optimizing the cost function, the projection algorithm is applied to deal with the unknown nonlinear terms, the RBFNN is used to compensate the effects of unknown actuator faults, and the anti-windup compensator is introduced to tackle actuator saturation. Theoretical analysis shows that the speed tracking errors of multiple subway trains are bounded, and the distance of adjacent subway trains is stabilized in a safe range. Finally, the validity of the proposed CMFAILC scheme is verified through multiple subway train simulations.