Stability-Enhanced Model Predictive Control for Urban Rail Transit Train

Abstract

Automatic train operation (ATO) control is a pivotal part of urban rail transit development, where designing the dynamics models and controllers for the ATO control scenarios presents a formidable challenge. To begin with, considering the fundamental resistances encountered by trains during the operation process, including elemental running resistance and time-varying slope resistance, we treat relative distance and relative speed between train carriages as state variables in the control modeling. Considering changes in traction/braking outputs as control variables, we formulate a meticulous dynamics model for urban rail transit trains (URTT). Furthermore, a stability-enhanced model predictive control (SEMPC) approach is proposed for ATO control in URTT, with a terminal term being added to the control objective design for stability requirements. This approach anticipates the future dynamic behaviors of the control system, yielding a stable and convergent predictive controller for the ATO system. Lastly, utilizing operational data from a specific urban rail line as an illustrative example, we conduct comparative analyses of the operational control performance among various controllers in scenarios of the single section, multi-section, and disturbance. Experimental results demonstrate that the proposed SEMPC controller exhibits superior performance to the compared controllers in terms of input cost, speed error, displacement error, and station-stopping error for ATO control in URTT.

Keywords

• Automatic train operation (ATO)
• urban rail transit train (URTT)
• stability-enhanced model predictive control (SEMPC)
• speed control