In response to the problems of complex interaction, poor path planning, and untimely information display in traditional roaming methods in the three-dimensional scenarios of rail transit engineering, an adaptive roaming method is proposed to solve the pain points of three-dimensional scene visualization and engineering information display and communication. By integrating long-line, large-scale, and multi-source heterogeneous data generated during rail transit informatization, building information modeling (BIM) and geographic information system (GIS) data fusion technology is used to build high-precision three-dimensional scenarios. Automatically plan to generate the optimal roaming path based on line geometric features and interest point distribution; combined with key technologies such as camera attitude dynamic control, roaming interest point capture, and information synchronization display, smooth transition of perspective and real-time presentation of key information are achieved. The technology is applied to the information construction process of one national railway and five intercity railways in Shenzhen, forming a rail transit BIM+GIS roaming system, providing an efficient and interactive three-dimensional visualization solution for rail transit project management, application display and communication.

The signaling system renovation project of Jinbin Light Rail Line 9 is facing strong constraints such as the type of of the terminal station for pre-station turnback, insufficient length of the train inspection garage line, and tight equipment room area. There are significant contradictions with the renovation goals of achieving a 2-minute turnback interval, upgrading the function of the automated train inspection garage, and installing equipment without disturbance. Targeted optimization design is needed. Firstly, the current research status of signaling system transformation was introduced, and the constraints of strong constraint conditions on design objectives were analyzed. Secondly, the focus of the design research is on three aspects: pre-station turnback capability, automated train inspection garage design, and signaling room equipment layout. The key bottlenecks that conventional design methods cannot meet the renovation goals under strong constraints are identified. Finally, this paper proposes a systematic optimization plan: adding axle counter equipment in the pre-station crossover section and changing the switch control logic from double-acting to single-acting, optimizing the safety braking model for train inspection garage stopping and adjusting the C-track interlocking logic, and adopting an uninterruptible power supply (UPS)/battery transition scheme for centralized stations while streamlining the equipment configuration at non-centralized stations. The application results show that the optimization measures effectively solve the contradiction between strong constraints and renovation goals. The turnback interval of the pre-station has been compressed from 136 seconds to within 120 seconds, meeting the requirement of 2-minute operation; The protection distance required for automatic stopping of trains has been shortened from 7.56 meters to within 3.8 meters, achieving automatic entry and exit of trains into the inspection garage; At the same time, the space requirements for equipment installation were reduced, and the equipment was installed without disturbance under the existing equipment room conditions.

This paper aims to effectively integrate knowledge from metro construction quality acceptance standards and address the current issues of singular knowledge organization methods, as well as isolated and fragmented knowledge. Based on the collection of metro construction quality acceptance standards issued by national and representative municipal authorities, generative AI was employed to extract key knowledge elements. A knowledge graph was then constructed using graph database software, and a multidimensional knowledge retrieval method along with node importance evaluation were proposed. Taking the construction of metro open-cut stations as an example, the proposed methods were applied to conduct an in-depth analysis of key knowledge at various stages of quality acceptance for this unit project, and acceptance recommendations were provided. [Conclusion] The knowledge graph of metro construction quality acceptance standards can offer comprehensive knowledge support for quality acceptance, providing significant reference value for improving the knowledge management level and utilization capability of metro construction quality acceptance standards.

Aiming at the common issues of insufficient scene adaptability and inadequate modeling of temporal feature dependencies in existing visual fatigue detection methods, this paper proposes an intelligent rail driver fatigue detection algorithm based on facial multi-features and Bi-LSTM. First, the algorithm uses Dlib to locate the driver's facial keypoints and calculates five fatigue feature parameters, including the Eye Aspect Ratio (EAR), Mouth Aspect Ratio (MAR), and head pose angles. It also proposes the frequency of longitudinal head scanning behavior for distraction detection. Second, the Bi-LSTM network is employed to establish temporal dependencies and complete the determination of the driving state. Third, temporal attention is introduced to enhance the representation capability of key features, improving the overall performance of the algorithm. Finally, the research results show that, compared to traditional models, the proposed method achieves an accuracy of 94.24% and improves the F1-Score from 0.8324 to 0.9027, validating the effectiveness and feasibility of the proposed approach. The findings can provide a reference for public transportation safety assurance.

In response to seismic events that render certain metro sections impassable or subject to speed restrictions, this study proposes an optimization model for metro train emergency rescheduling based on Earthquake Early Warning (EEW) information. The model proactively adjusts existing train operation plans before the earthquake occurs, upon receiving EEW alerts. It incorporates constraints such as arrival and departure time constraints, headway constraints, rolling stock circulation constraints, and earthquake impact-related constraints. The objective function aims to minimize the deviation between actual and ideal train headways, the number of skipped station stops, and the time trains operate within hazardous zones after the earthquake. A linearization method is used to convert the model into a Mixed-Integer Linear Programming (MILP) formulation. To verify the effectiveness of the proposed model under various earthquake scenarios, a case study is conducted using actual operational data from Line 6 of the Kunming Metro on a typical working day. The results show that the model can generate emergency train rescheduling plans based on EEW information under different earthquake intensities and affected areas. Different rescheduling strategies are adopted depending on the location of the train at the time of receiving the EEW alert. When speed restrictions are imposed due to the earthquake, a greater increase in inter-station running time leads to a more significant impact on train operations.

To address the need for efficient transportation of high-value-added and time-sensitive goods and resolve the challenge of selecting optimal freight paths within the complex urban rail transit networks of megacities, this study presents a comprehensive approach. First, a network simplification method is proposed to streamline the topological structure of urban rail systems by accounting for network complexity and uneven passenger flow distribution, thereby enhancing the computational efficiency of multimodal path optimization algorithms. Second, a path optimization model is developed with dual objectives of maximizing load rate and minimizing operational costs to balance cargo capacity utilization and economic feasibility. Finally, a hybrid algorithm integrating an improved ant colony optimization and particle swarm optimization is designed to solve the model, where particle swarm optimization optimizes the hyperparameters of the ant colony algorithm to accelerate convergence speed. Validated through a case study in Guangzhou, the methodology achieves significant performance improvements: network simplification reduces nodes by 76% and links by 69%, cutting algorithm runtime from 11,849 seconds to 1,613 seconds and boosting computational efficiency by 87.4%. The hybrid algorithm further reduces iterations from 46 to 15 while maintaining solution quality. Comparative analysis confirms the superiority of this approach in balancing computational speed, resource utilization, and path optimization effectiveness for urban rail freight systems.

To the lack of technologies and practical experience in non-suspension renovation of underground station platforms under non-reserved conditions in China, a study is conducted on the civil engineering renovation situation and the platform renovation practices of Tokyo Metro underground stations. Firstly, through statistics, the spatiotemporal distribution characteristics of Tokyo Metro stations that have undergone civil engineering renovation are analyzed, and the influence of laying method and reserved conditions on the timing and construction duration of renovation is explored. Subsequently, the process for civil engineering renovation of Tokyo Metro underground stations is delineated. Finally, based on specific cases, the main objectives and key points of platform extension, platform widening, platform addition and comprehensive platform renovation are analyzed, and relevant insights for the construction of metro stations in China are summarized. The study results show that about 15.1% of the current 259 underground stations of Tokyo Metro have undergone civil engineering renovation, which is a considerable number. The civil engineering renovation construction steps of underground stations usually follow “from outside to inside, build new first and then demolish old”. For the underground stations of Tokyo Metro, only platform extension may have reserved platform extension space based on the length of future train formation, whereas platform renovations under non-reserved conditions prioritize minimizing renovation scale, demonstrating that underground lines need to pay attention to reservation design with high cost-effectiveness and the underground station platform renovation under no reserved conditions needs to emphasize targeting. The long-term station renovation project needs to strengthen risk warning and dynamic control. Station-city integration requires balancing station passenger flow carrying capacity with surrounding land development intensity. The study findings can provide reference for the design and renovation of underground station platforms in China.

Tourism has become a major pillar of the national economy, and the integrated transport-tourism development model is steadily taking shape. Despite its inherent advantages, tourism rail transit still faces multiple constraints. This study therefore evaluates its current stage of development. Drawing on non-exhaustive data, it catalogs the principal completed, under-construction, and planned tourism rail transit projects nationwide. By the end of 2024, 23 projects had been completed, 10 were under construction, and 38 were in the planning pipeline; the 33 completed and under-construction lines together extend approximately 621.86 km, comprise 265 stations, and represent an aggregate investment of about 72 billion yuan. The analysis identifies four core challenges: unclear functional positioning, protracted and fragmented approval procedures, elevated construction costs, and underdeveloped technical standards and specifications. In response, the study advances targeted strategic recommendations. The findings offer a decision support reference for evidence-based planning, financing innovation, and the efficient implementation of tourism rail transit projects.

Addressing the critical challenges in constructing a data security protection system for intelligent rail transit, this paper systematically reviews recent research advances in this field. First, it examines the characteristics of rail transit data, including its multi-source heterogeneity, high real-time demands, and strong sensitivity, along with the primary security risks during data transmission, storage, and processing. Second, it explores the current application status, suitability conditions, and performance metrics of cutting-edge protection technologies—such as federated learning, lightweight encryption, and dynamic access control—against threats like cyberattacks and data breaches in resource-constrained rail transit environments. Finally, it identifies key future research directions, including AI-empowered active defense and cross-system collaborative governance. The findings demonstrate that federated learning effectively resolves data silo issues while enhancing model training efficiency under privacy guarantees. Lightweight encryption algorithms demonstrate millisecond-level response advantages in wayside equipment, exhibiting significant application potential and representing a major future development trend. The research outcomes provide both theoretical underpinnings and practical guidance for building an efficient and reliable intelligent rail transit data security protection system.

Existing evaluation approaches for urban rail transit predominantly focus on station-level analysis, lacking comprehensive consideration of the integrity and coordination at the corridor scale. This paper took Guangzhou's Metro Lines 2, 3, and 5 as empirical research objects, and extended the conventional node-place model from the station level to the corridor scale, aiming to better capture spatial dynamics along metro lines. The methods of Kernel Density Estimation and Corridor Balance Quantification were put forward to systematically assess the equilibrium status and development potential of stations along these corridors, and proposes tailored optimization strategies for each. The results reveal that Line 2 exhibited a “dependent” state overall, with both node and place values relatively low; Line 3 showed an “unbalanced node” state, featuring significant transportation advantages but underdeveloped place value; Line 5 approached a balanced state, with relatively coordinated development between node and place values. The study emphasized the necessity of integrated “corridor-station” dual-scale analysis to holistically evaluate future development needs and potentials along rail transit corridors. The results provided valuable references for formulating differentiated development strategies tailored to various rail transit corridors.