Iterative Learning Control for Automatic Train Operation with Discrete Gears

Abstract

In the traditional iterative learning control (ILC) for automatic train operation (ATO), control inputs are usually continuous signals. In this paper, a practical ILC is presented to carry out the train operation by discrete traction or braking force. The train motion dynamic model is described by linear time-varying perturbation model along with the reference trajectories, which can be identified by the historical data. The ILC based on the perturbation model can be easily used to the case with the continuous control signals because the updating law of the ILC can be derived theoretically. Then the proposed ILC method is extended to the case with discrete gears by transforming the ILC with discrete control signals into a well-defined mixed integer programming (MIP) problem. The proposed method has been illustrated on the simulation case. Simulation results show that the method can not only track the reference trajectories to a fine accuracy but also restrict the gear shift frequency of the operation process, which is helpful to improve the ride comfort index of the whole train operation.