Gu, Yiran (Beijing Jiaotong University), Wu, Yiling (Beijing Jiaotong University), Yan, Fei (Beijing Jiaotong University)

Integration of POMDP Planning under Uncertainty with MPC: An Application Study in Railway Virtual Coupling

Scheduled for presentation during the Invited Session "Control, Communication and Emerging Technologies in Smart Rail Systems II" (WeBT7), Wednesday, September 25, 2024, 14:30−14:50, Salon 15

2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC), September 24- 27, 2024, Edmonton, Canada

This information is tentative and subject to change. Compiled on October 7, 2024

Abstract

This paper proposes a train control method applied in railway virtual coupling technology. We simplify the virtual coupling formation into two trains, with the front train termed the leading train and the rear train termed the following train. When the state part of the leading train is observable, we use the Partial Observable Markov Decision Process (POMDP) model to predict the operational trajectory of the leading train. The prediction results are then relayed to the Model Predictive Control (MPC)-based controller of the following train, where MPC is utilized to control the following train and provide predictions for its operation. Experimental results indicate that such a method enables the following train to effectively track the leading train.