16 January 2024 in IEEE Transactions on Intelligent Transportation Systems
Trajectory prediction of the train ahead is vital for deciding the minimum safe distance for train separation. An important research challenge, in this regard, is how to effectively and accurately estimate the trajectory of the train ahead using spatiotemporal approaches with train operating data. In this paper, we present a novel hybrid model for predicting train trajectories by integrating Long Short-Term Memory (LSTM) and Kalman Filter (KF). In the proposed hybrid LSTM-KF model, the LSTM model is used to analyze time series data and discover the long-term dependencies of train trajectory data, which can be assumed as observation data for the KF model. The KF model can combine train dynamics mechanism to extract local features of train operation data, so as to smooth the train trajectory predicted by LSTM model. A novel on-the-fly algorithm is devised for efficient realization of the proposed LSTM-KF model. Experiments are carried out using the train data collected from Chengdu Metro Line 8. The results indicate that our hybrid LSTM-KF model has greatly improved the accuracy of long-term train trajectory prediction.
3 October 2024 in Vehicle System Dynamics
15 December 2025 in IEEE Transactions on Intelligent Transportation Systems
This study addresses the issue of communication link interruptions in train platoon control under the complex operating environment of high-speed railways. An adaptive trajectory tracking control approach based on a finite-time sliding mode is proposed under jointly connected switching topologies. The proposed framework ensures platoon consensus under jointly connected switching topologies, where the leader’s information need not reach all followers in each topology but only collectively over a finite interval, thus relaxing connectivity requirements and enhancing practicality. Within this framework, the sliding surface is designed to incorporate relative error signals determined by the communication topology, and an adaptive mechanism is employed to effectively handle unknown external disturbances without requiring prior knowledge of their bounds or derivatives. Simulation results demonstrate that, compared with MPC and non-adaptive approaches, the proposed strategy reduces position errors by approximately 80% and velocity errors by around 40%, significantly improving platoon tracking accuracy. Furthermore, simulations under both periodic and random topology switching indicate that periodic switching achieves performance closer to that of a fully connected topology, further enhancing trajectory tracking effectiveness.
2 July 2025 in IEEE Transactions on Intelligent Transportation Systems
Train platoons are a really innovative application for future railways, in which trains are virtually-coupled via Train-Train communication, so drastically reducing their headways and increasing line capacity. However, as the spacing between trains becomes closer, the influence of disturbances from preceding trains becomes more significant and cannot be ignored. This influence can have a substantial impact on the operational behavior of the following trains, leading to fluctuating spacing within the platoon and potentially compromising the safety and stability of the train spacing. This paper introduces a formalized modeling framework and a controller synthesis method for train platoons, aiming to achieve absolute safety and stable operation in train platoon operations. The framework utilizes Hybrid Priced Two-player Timed Game Automaton (HPTTGA) to describe train platoon dynamics and synthesizes controllers that satisfy safety requirements using an On-the-fly Controller Synthesis (OCS) algorithm, where the effectiveness of the algorithm is demonstrated through a two-train tracking example, and Q-learning is applied to select controller which satisfy local and string stability. Owing to model-based feature, the proposed method is compared to classical Model Predictive Control (MPC) in terms of safety, local stability, and string stability. Experimental results demonstrate that the proposed method consistently ensures train platoon safety and achieves similar control performance to MPC in extreme operating scenarios. Moreover, under velocity limit scenarios, the proposed method successfully achieves the desired following behavior of the trailing train according to the control objectives.
1 July 2026 in IEEE Internet of Things Journal