The rapid development of Artificial Intelligence (AI) has enabled powerful tools for processing large-scale data and extracting meaningful insights across multiple engineering fields. In railway engineering, AI techniques are increasingly being applied to infrastructure monitoring, with the goal of improving safety, reliability, and cost efficiency by detecting defects and predicting maintenance needs more effectively than traditional methods.
This thesis explores the integration of Physics-Informed models with Data-Driven approaches to monitor railway infrastructure. It aims to develop hybrid models capable of capturing both underlying physical behaviours and complex data-driven patterns. To this end, knowledge of train-track interaction will be combined with machine learning algorithms. In the end, the study will assess the potential of such approaches to enhance anomaly detection and support more effective predictive maintenance strategies.
Contacts: Ivano La Paglia (ivano.lapaglia@polimi.it); Egidio Di Gialleonardo (egidio.digialleonardo@polimi.it); Alan Facchinetti (alan.facchinetti@polimi.it)