Long and Short-Term Memory Mechanism Hybrid Model for Speed Trajectory Prediction of Heavy Haul Trains*

The trajectory prediction of heavy heavy trains (HHTs) is crucial for ensuring the safe and automatic operation. It is inevitable to design a train model for predicting train operation trajectory of the HHTs. However, large capacity and the existence of unmodeled dynamics in the operation process caused by air resistance, working condition switching, and external environment make it difficult to establish accurate speed trajectory prediction models (STPMs) using traditional mechanism-driven methods. Most recent research considers using data-driven model to learn information from data, but they have no information about the physical characteristic of the STMP model. This makes it difficult to accurately describe the relationship between the control force and the running speed during the train operation. To overcome these issues, this study combines the mechanism-driven model with the deep learning model to establish a new long and short-term memory mechanism hybrid (LSTMMH) model. Specifically, the mechanism model describes the change of control force, while the LSTM captures the unmodeled dynamics of long time series of running process. The effectiveness of the proposed method is demonstrated while the performance is compared with the traditional LSTM and mechanism models using the real data.