Urban rail networks rely on infrastructure most passengers never see, from tunnels and ventilation shafts to the soil and road layers sitting above them. That makes gradual ground movement difficult to detect, particularly when early changes are too small to produce obvious surface damage.
Researchers at Seoul National University of Science and Technology (SEOULTECH) have tested a monitoring framework that combines satellite-based Interferometric Synthetic Aperture Radar (InSAR), terrestrial laser scanning and ground penetrating radar (GPR). The aim is to give infrastructure operators a way to screen large areas first, then concentrate more detailed investigation on locations showing unusual movement.
Led by civil engineering master’s student Tae-Yong Park, the research team applied the method to more than 16 kilometers of Seoul’s Bundang Line between Suseo Station and Cheongnyangni Station. The study was published in Tunnelling and Underground Space Technology on August 1 after appearing online in April.
Layered monitoring narrows the search for settlement
Ground settlement can develop for several reasons, including excavation, changing soil conditions, underground construction and inadequate compaction. In a dense city, determining where that movement is happening — and whether it warrants intervention — can require a mix of monitoring methods.
The SEOULTECH approach assigns a different task to each technology rather than relying on a single dataset.
InSAR provides the widest view. By comparing repeated satellite radar observations, researchers can track changes in surface elevation over time across a relatively large area. The team also used seasonal-trend decomposition with LOESS to separate longer-term settlement patterns from recurring seasonal movement.
Along the Bundang Line, the analysis highlighted a ventilation shaft with a developing settlement pattern. The satellite result was not treated as proof of structural failure. Instead, it gave researchers a location where closer inspection could be justified.
That distinction is important for transit operators. Conducting detailed surveys across every tunnel, shaft and road section would require considerable time and resources. Satellite monitoring could instead act as an initial screening layer, helping maintenance teams decide where higher-resolution investigation is most useful.
Researchers then examined the selected ventilation shaft using terrestrial laser scanning. The technology produced a detailed 3D point cloud that allowed deformation across the structure to be measured.
Visual inspection had already identified cracks and signs of previous repair work in the ceiling. Laser scanning showed that settlement increased toward the section of the shaft located directly below the road surface.
Those findings informed the next stage: a GPR survey of the roadway. Ground penetrating radar sends electromagnetic signals below the surface to identify changes in materials, structures and layer boundaries.
The survey indicated void-like features near the ventilation shaft and irregularities in subsurface boundaries. Together, the results pointed to uneven conditions in the ground surrounding the structure.
Using the technologies in sequence allowed the team to move from city-scale monitoring to site-level investigation. Satellite data identified the broader movement signal, laser scanning measured structural deformation, and GPR provided additional information about what was happening below the road.
Cross-checking different datasets could also help operators avoid overreacting to an anomaly caused by the limitations of a single sensing method.
Targeted inspections could support preventive maintenance
The operational value of the framework is less about replacing existing inspections and more about deciding where to deploy them.
Urban transit agencies manage large networks of infrastructure that cannot realistically be surveyed continuously at the same level of detail. A multi-stage monitoring system could help prioritize assets showing unusual movement, directing engineering teams toward locations where further investigation may be warranted.
That could support a more preventive maintenance model. Instead of waiting for settlement to become visible through pavement deformation, cracking, sinkholes or service disruption, operators could use changes detected over time to trigger earlier assessment.
The approach may also be relevant as cities add new tunnels, utilities and developments around existing rail systems. Construction activity above and below ground can alter local conditions, while older infrastructure may be exposed to stresses that were not present when it was built.
However, the study covered one section of one subway network. Its performance still needs to be assessed across different geological conditions, infrastructure types and urban environments.
InSAR measurements can be influenced by atmospheric effects, satellite observation geometry and the number of suitable radar targets available in an area. Laser scanning and GPR also require specialist interpretation, and a detected anomaly does not automatically indicate an immediate safety problem.
For operators, the framework is therefore better considered an additional source of infrastructure data rather than a substitute for geotechnical analysis, structural inspection or established maintenance programs.
Even with those limitations, the research demonstrates how combining sensing technologies at different scales could make underground monitoring more selective. For transit agencies facing long asset lists and limited inspection capacity, knowing where to look more closely may be just as valuable as collecting more data.