Selected work

Representative Papers

A Multi-Sensor Dataset for Monitoring the Operational Environment of Rail Vehicles

Aug 05, 2026

This work addresses the critical lack of high-quality, large-scale, multi-sensor annotated data tailored to varying levels of railway automation (GoA2–GoA4), which has hindered the safe and efficient deployment of AI-driven systems in rail environments. To bridge this gap, the authors construct and publicly release a multimodal dataset specifically designed for railway operational scenarios, integrating synchronized data from visible-light cameras, LiDAR, and other sensors across diverse operating conditions. Through a rigorous, high-precision manual annotation pipeline, the dataset comprises over 7 million labeled instances, enabling robust detection and classification of both rail-specific and general objects. This resource fills a significant void in training and validation data for fully automated rail systems, substantially advancing the development, evaluation, and real-world deployment of AI models in railway applications.

0 citationsRead paper

A GitOps-Driven Annotation Catalog for Fully Automatic Railway Operations

Aug 05, 2026

This work addresses the challenges of managing high-quality, dynamically evolving annotated data essential for Grade of Automation 3–4 (GoA3–GoA4) train operations, where existing data catalogs suffer from high operational overhead, weak development integration, and documentation drift. To overcome these limitations, the authors propose a lightweight, GitOps-driven data catalog architecture that introduces the Data-as-Code paradigm into railway AI perception systems for the first time. By leveraging CI/CD pipelines and static site generation, the approach enables versioned data management, end-to-end traceability, and automated documentation. This solution significantly reduces operational costs, enhances development integration, ensures regulatory compliance, and automatically produces high-performance dataset overview pages.

0 citationsRead paper

A System for Train Condition Monitoring and Structural Health Assessment of Rail Vehicles

Aug 05, 2026

This study addresses the challenges of achieving Grade of Automation 4 (GoA4) in mainline railway operations, where complex open environments and stringent safety requirements hinder reliable detection of train collisions, run-overs, and structural damage using conventional perception systems. To overcome these limitations, this work proposes an innovative framework that integrates structural sensors with artificial intelligence to enable real-time vehicle state monitoring and impact identification. For the first time, the approach achieves accurate online recognition of such critical events alongside continuous assessment of structural health. The proposed method not only enhances operational safety but also facilitates predictive maintenance and long-term vehicle design optimization, thereby providing essential technical support for the deployment of fully automated railway systems.

0 citationsRead paper

Driving-Over Detection in the Railway Environment

Feb 19, 2026

This study addresses the lack of reliable automated detection methods for wheel-rail crushing events—such as wheels running over foreign objects on tracks—in fully automated train operations. It presents the first systematic investigation into automatic recognition of such events in railway scenarios. A dataset was constructed through controlled crushing experiments involving objects of multiple materials, and the authors propose and compare a convolutional neural network (CNN)-based detection approach against two classical thresholding methods. Experimental results demonstrate that the proposed CNN model achieves an average accuracy of 99.6%, substantially outperforming the threshold-based methods, which attain accuracies of 85.0% and 88.6%, respectively. These findings confirm the superiority and feasibility of deep learning for this critical railway safety task.

0 citationsRead paper

Recent publications

Latest Papers

A Multi-Sensor Dataset for Monitoring the Operational Environment of Rail Vehicles

Aug 05, 2026

This work addresses the critical lack of high-quality, large-scale, multi-sensor annotated data tailored to varying levels of railway automation (GoA2–GoA4), which has hindered the safe and efficient deployment of AI-driven systems in rail environments. To bridge this gap, the authors construct and publicly release a multimodal dataset specifically designed for railway operational scenarios, integrating synchronized data from visible-light cameras, LiDAR, and other sensors across diverse operating conditions. Through a rigorous, high-precision manual annotation pipeline, the dataset comprises over 7 million labeled instances, enabling robust detection and classification of both rail-specific and general objects. This resource fills a significant void in training and validation data for fully automated rail systems, substantially advancing the development, evaluation, and real-world deployment of AI models in railway applications.

0 citationsRead paper

A GitOps-Driven Annotation Catalog for Fully Automatic Railway Operations

Aug 05, 2026

This work addresses the challenges of managing high-quality, dynamically evolving annotated data essential for Grade of Automation 3–4 (GoA3–GoA4) train operations, where existing data catalogs suffer from high operational overhead, weak development integration, and documentation drift. To overcome these limitations, the authors propose a lightweight, GitOps-driven data catalog architecture that introduces the Data-as-Code paradigm into railway AI perception systems for the first time. By leveraging CI/CD pipelines and static site generation, the approach enables versioned data management, end-to-end traceability, and automated documentation. This solution significantly reduces operational costs, enhances development integration, ensures regulatory compliance, and automatically produces high-performance dataset overview pages.

0 citationsRead paper

A System for Train Condition Monitoring and Structural Health Assessment of Rail Vehicles

Aug 05, 2026

This study addresses the challenges of achieving Grade of Automation 4 (GoA4) in mainline railway operations, where complex open environments and stringent safety requirements hinder reliable detection of train collisions, run-overs, and structural damage using conventional perception systems. To overcome these limitations, this work proposes an innovative framework that integrates structural sensors with artificial intelligence to enable real-time vehicle state monitoring and impact identification. For the first time, the approach achieves accurate online recognition of such critical events alongside continuous assessment of structural health. The proposed method not only enhances operational safety but also facilitates predictive maintenance and long-term vehicle design optimization, thereby providing essential technical support for the deployment of fully automated railway systems.

0 citationsRead paper

Driving-Over Detection in the Railway Environment

Feb 19, 2026

This study addresses the lack of reliable automated detection methods for wheel-rail crushing events—such as wheels running over foreign objects on tracks—in fully automated train operations. It presents the first systematic investigation into automatic recognition of such events in railway scenarios. A dataset was constructed through controlled crushing experiments involving objects of multiple materials, and the authors propose and compare a convolutional neural network (CNN)-based detection approach against two classical thresholding methods. Experimental results demonstrate that the proposed CNN model achieves an average accuracy of 99.6%, substantially outperforming the threshold-based methods, which attain accuracies of 85.0% and 88.6%, respectively. These findings confirm the superiority and feasibility of deep learning for this critical railway safety task.

0 citationsRead paper