Área: Engenharia Informática e de Computadores

Despacho de nomeação de Júri

Edital

Título da Tese: Actionable Spatiotemporal Pattern Analysis in Transport Engineering Systems: Tackling Open Challenges

Local da Prova:  Anfiteatro PA-3 (Piso -1 do Pavilhão de Matemática) do IST 

Data: 11/05/2026

Hora: 14h00
Abstract: With the increasing sensorisation of transport systems, high volumes of spatio-temporal data are being generated, creating unprecedented opportunities for monitoring multimodal end-to-end traffic flows (from origin to final destination) and subsequently supporting operational and tactical planning. Leveraging these data offers the potential to address core challenges associated with enhancing service quality, improving operational efficiency, and advancing the sustainability of mobility systems and their surrounding environments. In this context, the identification of multivariate indicators capable of representing end-to-end movements, with context-aware integration and guided by statistical and artificial intelligence approaches for mapping vulnerabilities in transport systems, plays a decisive role in extracting actionable knowledge and supporting evidence-based decision-making processes. However, significant challenges persist. The incompleteness of available data limits the robust assessment of operational performance metrics for the end-to-end movements. Additionally, the lack of systematic integration of contextual data (such as demographic information, land use, or modal attractiveness) results in a partial view of operational vulnerabilities and the pursuit of sustainability objectives. Finally, the high dimensionality inherent to multivariate data structures with fine spatiotemporal resolutions hinders the extraction of interpretable patterns that enable understanding of systemic transport dynamics. This thesis addresses these challenges through the development of data-driven methodological solutions for the monitoring and analysis of origin–destination operations in passenger transport systems. The proposed methodology advances the state-of-the-art by: (i) inferring and validating alighting information to reconstruct complete multimodal passenger trajectories; (ii) deriving performance indicators to characterise system operations; (iii) integrating contextual data; and (iv) modelling of performance indicators as multivariate time series for the extraction of actionable patterns. The validation of the developed contributions focuses on the multimodal passenger transport system of the Lisbon Metropolitan Area. The proposed methods enable an integrated analysis of OD corridors, supporting transport system optimisation, particularly in addressing competitiveness and sustainability concerns.