To enhance the train transportation efficiency under this system and meet the operational needs of dynamic marshalling adjustment, an online marshalling technology for urban rail train control systems based on vehicle-to-vehicle (V2V) communication is proposed, including flexible marshalling and virtual marshalling technologies. Firstly, the technical scheme and operation procedures for V2V-based virtual marshalling are established, and the core technologies under flexible marshalling are expounded, such as the setting of marshalling area parking points, determination of the vehicle-mounted control end, verification and protection of marshalling forms, collision envelope schemes, and vehicle- vehicle-mounted integration technology. Secondly, the technical scheme and operation procedures for the virtual marshalling function of V2V communication are established, and the core technologies under virtual marshalling are expounded, such as high-performance V2V communication technology, V2V communication security protection technology, collaborative control technology, multi-sensor fusion technology, etc. The results show that compared with the traditional CBTC marshalling technology, the the V2V-based marshalling technology reduces operational time by approximately 14 seconds and improves efficiency by about 34%. This study aims to provide a theoretical basis and reference for the marshalling technology operations of the next-generation train control system.

To investigate how urban express lines can achieve "rapid arrival" and enhance their service level and public transit attractiveness, this paper uses Chengdu Rail Transit Line 19 as a case study to examine key technologies in transportation organization oriented toward rapid arrival. Under Chengdu's dual-airport configuration of "one city, two airports," the Chengdu Rail Transit Line 19, which achieves "30-minute rapid connectivity between dual airports," represents an exemplary case of rapid arrival for urban express lines. Facing a 66.7 km operating distance between the two airports and a composite functional positioning, Chengdu Rail Transit Line 19 innovatively adopts the following key train operation technologies to achieve rapid inter-airport connectivity: Completing the design of a "one-line separation and three-line through operation" scheme among Lines 17, 18, and 19 to ensure the feasibility of direct train service between the two airports; increasing the maximum operating speed of the urban AC A-type trains from 140 km/h to 160 km/h, thereby creating better conditions for 30-minute direct connectivity between the dual airports; maintaining high-speed operation of direct trains by increasing their passing speed through stations to 160 km/h and introducing No. 42 turnouts. Through a systematic review of these innovative technologies, this paper identifies five key design insights for achieving rapid arrival in urban express line transportation organization: Advance planning to ensure cross-line operation capabilities; unified scheduling of cross-line operating lines; new approaches to speed enhancement for urban A-type vehicles; favorable conditions for increasing express train passing speeds; and cautious introduction of No. 42 turnouts into underground projects.

To promote the coordinated development of urban rail transit with urban spatial layout, population and employment distribution, and public service facilities, this paper constructs a multi-dimensional (public transport accessibility level, PTAL) quantification model that integrates urban rail transit and conventional bus services based on a comprehensive urban transportation model. This model is used to systematically evaluate the public transport accessibility within the Dongguan metropolitan area. The study reveals that the current PTAL level in Dongguan is generally low, with average accessibility index ratios for the central urban area, sub-centers, economically strong towns, and ordinary towns recorded at 3.6:0.8:1.2:1. This indicates a spatial pattern characterized by "single-core polarization, small-scale decentralized clusters in peripheral areas, and homogeneous fragmentation in the remaining regions." The findings highlight several issues, including a disconnect between public transport resource allocation and the polycentric structure, a lack of synergy between urban rail transit and bus services, and a mismatch between public transport resource provision and the distribution of population and employment. The research demonstrates that by establishing a network of urban rail transit express lines and strengthening public transport services in underdeveloped areas, the overall public transport accessibility in the metropolitan area could increase by 65.3%. Notably, sub-centers such as Songshan Lake and Binhaiwan could experience an improvement of up to 153.6%, significantly reinforcing the polycentric structure defined as "one primary center, two sub-centers, and six clusters." Furthermore, optimizing bus connections at urban rail transit stations and coordinating the layout of public facilities can enhance the alignment between rail transit resources and the distribution of population and public services. Ultimately, this study provides technical support for establishing an integrated "rail transit+bus" system to advance polycentric development in Dongguan.

In response to ambiguities and inconsistencies in China's evaluation standards for train-induced environmental vibrations in urban rail transit, which affect assessment conclusions, this paper adopts a combination of field measurements and statistical analysis to conduct a comparative study of current standards regarding evaluation indicators, weighting curves, analysis frequency ranges, and determination of train pass-by intervals. The main conclusions are as follows: it is recommended to establish both a time-frequency comprehensive evaluation indicator and a frequency-domain evaluation indicator. It is suggested to set the single-value evaluation indicator as the maximum vertical vibration level (VLz_max), the frequency-domain octave evaluation indicator as the maximum vibration level of 1/3 octave bands (VAL_max), and adopt the peak value retention method in the calculation. The W_k weighting curve is suggested to be adopted and the analysis frequency range is recommended as 1~80 Hz. To effectively eliminate the influence of subjective factors in determining train passing periods on evaluation results, it is recommended to use time weighting for vibration signal analysis and the peak hold method for calculating VLz_max and VAL_max.

In view of the limitations of the two-dimensional perspective of the traditional node-place model in complex urban environments, this study constructs a three-dimensional "node-place-vitality" model. With 41 stations of Jinan's rail transit as the research subjects, a service radius of 800 meters around each station is demarcated. Twenty-four indicators are selected from three dimensions: traffic function (node), urban space (place), and dynamic vitality (vitality). The comprehensive subjective-objective weighting method and K-means clustering analysis are employed to evaluate the coordination of station areas. The results indicate that the three-dimensional model can effectively uncover the synergistic relationships among nodes, places, and vitality. Moreover, the stations are classified into five types: high-vitality hub stations, vitality-improvement stations, balanced-development stations, node-reinforcement stations, and comprehensive-cultivation stations. Their spatial distribution is highly associated with the development characteristics of Jinan's main urban centers, sub-centers, and marginal areas. This research offers a scientific foundation for the optimization of Jinan's rail transit station areas and the formulation of TOD strategies. Additionally, it provides a novel perspective for studies on the coordination of station areas in similar cities.

To address the critical issue of predicting the time-series passenger flow at subway stations, a novel hybrid deep learning method, AFT-K-GRU⁺, is proposed to provide more accurate decision support for subway operation and management. Leveraging the multi-scale temporal dependencies inherent in subway passenger flow data, the model innovatively integrates three core components: the attention-free transformer (AFT), the gated recurrent unit (GRU), and the KAN network. Specifically, the AFT module captures long-term periodic patterns in passenger flow data, the GRU module focuses on extracting short-term fluctuation characteristics, and the KAN network enhances the model's expressive capacity. To enhance the robustness of the model's predictions, this study innovatively designs a knowledge-driven correction controller based on residual feedback, which enables online self-correction of the output. Experimental validation using real-world card swipe data from Beijing subway stations demonstrates that the AFT-K-GRU⁺ model effectively predicts station passenger flow, achieving mean error reductions of 25.3% and 18.7% compared to Transformer and LSTM methods, respectively. This model offers effective methodological support for subway operators, assisting management departments in passenger flow control and daily planning, thereby enhancing the operational efficiency and service quality of subway lines.

The horizontal alignment design of urban metro shield sections is prone to being constrained by dense structures and land use, which usually results in increased investment and demolition. An optimization model for dual horizontal alignments in metro shield tunnel sections is proposed, which takes the intersection points of the left line and the right line as independent variables. The specification requirements are transformed as geometric constraints, and penalty terms for land use under-passing and line spacing violations are introduced in the objective function. The model is solved by a collaborative differential evolution algorithm with separate populations for the left and right lines. The case study shows that the proposed model can optimize a 2.7 km horizontal alignment in approximately 150 seconds, reducing the total investment by approximately 0.3%. It automatically avoids control areas and guides the automatic parallel or separated design of horizontal dual lines in shield tunnel sections.

To address the high risks faced by metro systems due to their enclosed underground spaces, vulnerable facilities, and structural complexity amidst frequent rainstorm waterlogging, it is imperative to develop a scientific vulnerability assessment method for metro stations to enhance comprehensive resilience. Based on the chain analysis framework of the driver-pressure-state-response-control (DPSRC) model, this study systematically analyzes the dynamic causality of metro rainstorm waterlogging risks and constructs an assessment index system comprising 5 criterion layers and 20 indicator layers. By integrating subjective and objective weights using the AHP-entropy weight method, a comprehensive vulnerability assessment approach for metro stations was proposed and applied to an empirical analysis of 99 stations in the Ningbo Metro. The results indicate that: ① The proposed method can accurately and effectively assess station vulnerability; ② Stations vulnerabilities exhibit an asymmetric distribution across high (19.10%), medium (51.69%), and low (29.21%) vulnerability levels, with 70.79% of stations exposed to risk, highlighting an urgent need for systemic prevention and control; ③ Waterlogging prevention capacity, monitoring and early warning capabilities, and emergency equipment reserves constitute the key contributing factors to high-vulnerability stations. This research provides a reference for rainstorm waterlogging risk management, disaster prevention, mitigation planning, and resilient transportation development in urban metro systems.

In drill-and-blast tunnel construction, the geometric configuration and retained dimensions of the core soil directly influence tunnel settlement magnitude and surrounding rock stability. Traditional manual monitoring methods are inherently subjective, inefficient, and fail to meet real-time monitoring requirements. This paper proposes an automated core soil monitoring method based on an improved YOLO framework. By defining core soil feature points, a standardized annotation system is established. A lightweight detection model incorporating the Convolutional Block Attention Module (CBAM) is designed, effectively suppressing false detections caused by uneven illumination, dust interference, and blurred soil boundaries. Integrated with perspective geometry vanishing point prediction techniques, the VPTR model is employed for vanishing point detection and geometric constraints. This approach corrects perspective distortion errors in core soil cross-section shape calculation, incorporates core soil dimensions and proportions into potential risk classification criteria. For engineering deployment, knowledge distillation and dynamic precision-switching techniques are adopted to enable lightweight operation on edge computing devices, supporting the real-time decision-making demands of "monitoring while excavating".

To address the challenge of managing excessive deformation defects in operational shield tunnels in China, a comprehensive tunnel deformation correction technology system centered on the Radial Cement-Soil Pile Method for Correcting Excessive Tunnel Deformation has been developed. This system actively constructs a spatial network reinforcement body composed of horizontal, vertical, and inclined cement-soil piles in the soil outside the tunnel. It formulates an optimized construction sequence of "first constructing horizontal piles to restrict convergence, then constructing vertical piles to lift settlements, and finally constructing inclined piles to coordinate deformation" and adopts refined strategies such as graded pressure control to accurately apply correction forces. Meanwhile, based on the principle of minimum potential energy, a semi-empirical formula for predicting the single-ring lifting and convergence correction values has been derived, thereby forming a complete set of technical measures. This technology system has been applied and verified in a metro tunnel section in Northeast China. The results show that the Radial Cement-Soil Pile Method for Correcting Excessive Tunnel Deformation can synergistically correct settlement and convergence issues, effectively stabilize the post-construction settlement rate, and significantly improve leakage defects. The research results provide an innovative technical solution for the treatment of defects in operational shield tunnels and have reference value for similar engineering projects.