A comprehensive review of railway track condition monitoring: From sensor selection to data processing for defect-oriented, task-driven applications

This comprehensive review examines recent advances in railway track condition monitoring, summarizing state-of-the-art sensor technologies and data-driven analytical approaches. Representative monitoring sensors and devices are discussed to highlight innovations in combined large-scale and local sensor deployment strategies and non-destructive testing, with particular emphasis on vulnerable track sections such as transition zones and switches, as well as key challenges in practice, including environmental stressors and stringent operational demands. The review of data processing approaches evaluates the advantages and limitations of different data types and identifies persistent gaps, notably the lack of standardized track health indices and the limited availability of open datasets. Previous reviews have primarily focused on the advancement of specific sensor types or data processing techniques; however, they often address these aspects in isolation, resulting in a fragmented perspective that lacks an integrated sensor-data lifecycle framework. This limitation is particularly critical given the increasing diversity of sensor technologies and the growing volume and heterogeneity of data, which make it challenging for practitioners to select appropriate sensors and analytical methods for specific railway track condition monitoring tasks. By synthesizing global research trends, this work provides a structured roadmap for advancing defect-oriented, task-driven applications to support predictive maintenance practices in modern railway infrastructure.