🤖 AI Summary
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.
📝 Abstract
Automatic train operation (ATO) at grade of automation 3 and above (GoA3-GoA4) requires robust AI-based perception systems capable of reliably detecting obstacles and railway-specific objects under real-world conditions. The effectiveness of these modern artificial intelligence approaches depends heavily on large-scale, high-quality, and highly dynamic annotated datasets. However, managing metadata, maintaining provenance, and tracking the iterative evolution of these annotations impose significant infrastructural and regulatory requirements. Existing monolithic data catalogs often suffer from massive operational overhead, poor integration into developer workflows, and severe documentation drift. This paper introduces an innovative, lightweight GitOps-based architecture for metadata management. By leveraging Data-as-Code principles, Continuous Integration/Continuous Deployment (CI/CD) pipelines, and Static Site Generation (SSG), the proposed approach establishes a seamless, developer-centric workflow. This ensures an traceability, enforces strict regulatory compliance, and automatically generates a highly performant dataset overview.