3D point cloud semantic segmentation plays a significant role in scene understanding and has been widely applied in various downstream tasks, such as autonomous driving and augmented reality. In this paper, we focus on point cloud segmentation on railway scenarios that remains underexplored in previous studies. First, we meticulously construct a railway point cloud dataset, Railway360, which is collected from a main railway line in China, covering 4.1 kilometers and approximately 800 million points divided into 10 categories. Thanks to the usage of a single-line high-frequency LiDAR and the specific vertical scanning pattern, Railway360 provides not only the position coordinates, intensities, and category labels, but also the timestamps of the point cloud collection. This allows us to obtain a continuously complete and uniformly sampled point cloud dataset, enabling more accurate temporal analysis of the scene, which can largely benefit subsequent tasks such as scene segmentation and track equipment measurement. Second, we propose a novel Mamba-based framework, named RailSeg-Mamba, for railway 3D segmentation, where spatial and temporal features are fully integrated by leveraging the significant capabilities of modeling the sequence features of the selective state space models. To our knowledge, this is the first framework that takes the temporal features of point cloud into account. We extensively evaluate the performance of the state-of-the-art methods and provide a new benchmark on our dataset. The results demonstrate that modeling point clouds as sequences provides a promising solution for railway segmentation. The code and dataset are available at https://github.com/Assassincup/RailwaySegmentation.
Railway turnout (RT) equipment is an important component of railway signaling systems, and its reliable operation is essential for railway safety. To improve maintenance efficiency and ensure equipment reliability, this study proposes a fault prediction method based on RT action current curves collected by the microcomputer monitoring system, combining Sparse Autoencoder (SAE), Squeeze and Excitation Networks (SE), and Bidirectional Long Short-Term Memory (Bi-LSTM). Based on the 7 fault and 1 normal modes, the pre-fault degradation current data of the ZD6 RT were collected in chronological order, and the labels were assigned according to the fault types subsequently confirmed by field maintenance records. Secondly, an RT fault prediction model based on SAE fusion SE and Bi-LSTM was constructed. Because hyperparameters directly affect the performance of the prediction model, Adam algorithm is used to optimize the model parameters, and the Bayesian Optimization (BO) is used to optimize the hyperparameters of the RT fault prediction model. The proposed model is designed to learn the stage-dependent characteristics and sequential relationships embedded in each complete turnout action current curve and to predict the corresponding fault state and fault type. Finally, the experiment proved that this method can accurately predict the fault status of the RT, so that on-site maintenance personnel can predict the fault in advance and take maintenance measures, improve the maintenance efficiency of the RT and ensure the safe operation of the train.
Future high-speed railway systems rely heavily on secure and efficient wireless communication for train operation, passenger services, positioning, and safety-related functions. Antenna systems play a critical role in establishing these communication links; however, their design and implementation in high-speed rail environments present unique challenges compared to conventional vehicular and terrestrial platforms. Factors such as high mobility, large metallic train structures, aerodynamic constraints, harsh propagation conditions, and stringent reliability requirements impose significant demands on antenna integration and performance. This paper provides a comprehensive overview of antenna systems for high-speed railways from both engineering and deployment perspectives. Particular emphasis is placed on implementation challenges, including train-mounted and trackside antennas, with detailed discussion on practical integration aspects such as platform effects, radome design, electromagnetic compatibility, and maintenance. Key propagation scenarios in high-speed railway operations—such as open track, tunnels, viaducts, and stations—are analyzed in terms of their impact on antenna radiation behavior and overall system performance. Furthermore, state-of-the-art antenna designs are systematically classified based on radiation characteristics, frequency bands, and functional roles, ranging from omnidirectional and multi-antenna systems to dual-band and emerging high-frequency solutions. Beyond performance metrics, this review also examines validation methodologies, comparing simulation-based studies with real-world measurements from operational railway environments. Finally, future challenges and research directions are outlined in the context of next-generation railway communication standards and emerging technologies. Overall, this review aims to bridge the gap between theory and practice by providing practical insights for researchers and railway engineers involved in the design, integration, and deployment of antenna systems for high-speed railways.
Virtual Coupling (VC) aims to reduce inter-train following distances by relying on relative braking and vehicle-to-vehicle information exchange. Although VC train-following control is commonly addressed using model-based approaches such as Model Predictive Control, maintaining safe operation under communication delay, adhesion degradation, and positioning uncertainty remains challenging. This study proposes a safety-shielded Deep Reinforcement Learning (DRL) architecture for VC train following. A Soft Actor-Critic policy generates continuous nominal control commands for speed tracking, headway regulation, comfort, and actuation effort, while a braking-distance-based safety shield supervises the hard minimum-distance constraint through a one-step action-projection mechanism whenever the projection remains feasible. Therefore, the DRL policy is not treated as a standalone safety-certifying controller; safety is provided by the combined policy–shield architecture. The proposed method is evaluated under nominal operation, 10% adhesion loss, a communication-delay sweep (τ ∈ {0.0, 0.5, 1.0, 2.0, 5.0, 10.0} s), and a combined-uncertainty stress test involving delay, adhesion loss, and position noise. All scenarios are assessed over 50 randomized seeds using worst-case and distributional statistics. Additional diagnostics, including safety-shield activation, raw-action feasibility, correction magnitude, and shield-off evaluation, are used to distinguish the role of the learned policy from that of the safety shield. The results show safe operation under nominal conditions, 10% adhesion loss, and communication delays up to 2.0s in the considered setting. In contrast, delays of 5.0s and 10.0s and the combined-uncertainty stress test lead to safety violations and early termination, identifying the feasibility boundary of the considered VC formulation. Overall, the study provides a reproducible safety-shielded DRL baseline for VC train following and clarifies the delay regimes under which the proposed architecture remains applicable.
This study investigates the heat of hydration and internal temperature field of concrete in the precast construction stage of the 32m box girder for high-speed rail. The aim is to guide early thermal insulation and maintenance practices for the box girder, aiming to prevent temperature cracks or internal concrete damage resulting from significant temperature differentials or rapid cooling rates. Through testing, the adiabatic temperature rise of concrete was measured. By combining numerical analysis and actual measurements, this study analyzed the variations in the temperature field during the precast stage of the box girder and fitted an exponential function to represent the internal temperature variations over time. Results indicate that the internal temperature of the box girder concrete sharply increases from 5 to 15hours after casting, peaking between 15 to 20hours at around 61.0-61.3°C with a maximum rise of approximately 41.3°C. Furthermore, it was observed that the internal heat dissipation of concrete in the web of the box girder is inefficient, with a slow cooling rate averaging about 0.209°C/h. To mitigate temperature differentials, it is recommended to provide adequate insulation, especially during extreme cooling conditions or winter construction. This study highlights the significance of the 20-60hours period post-casting as a critical phase for curing, necessitating appropriate measures to prevent temperature-related cracks due to internal/external temperature disparities. Finally, the research findings demonstrate that the internal temperature variations in various sections of the box girder adhere to an exponential function, with MATLAB-developed models offering improved predictions for guiding construction activities and thermal insulation maintenance. These research results offer a scientific foundation for enhancing precast construction methods and maintenance practices, ultimately reducing the risk of early concrete cracking.
