National Technical University of Athens (NTUA)
A multilevel integrated assessment of safe and green mobility
Abstract
dc:descriptionThe main objective of this doctoral dissertation is the multilevel integrated assessment of safe and green mobility. This assessment is conducted across multiple levels, from individual trips to the road network, by fusing road infrastructure, traffic, and weather-related data with high-resolution naturalistic driving behavior data. To achieve this objective, the methodological framework is structured around three complementary analytical modules with distinct but interrelated objectives. The first analytical module focuses on the identification of safe and green mobility patterns at the trip and spatial levels. At the trip level, six sustainable profiles revealed structured heterogeneity in the integrated consideration of surrogate driving safety and fuel consumption, ranging from the most sustainable, namely "Low-risk x Eco", to the least sustainable, "Risky x Inefficient". These profiles were predicted using Machine Learning classifiers based on driving, infrastructure, weather, and temporal predictors. The strong performance of the XGBoost model confirmed the stability of the identified trip patterns, while SHapley Additive exPlanations (SHAP) analysis highlighted average driving speed, speed variability, stopping behavior, and route complexity, in terms of road grade and road junction exposure, as the most influential determinants. At the spatial level, 182 crash risk hotspots and 232 fuel consumption hotspots were identified within the study area through the application of global and local spatial autocorrelation statistics. Limited direct overlap was observed, indicating that the most crash-prone junctions are not necessarily the most fuel-inefficient. However, statistically significant short-range spatial segregation showed that safety and environmental mobility externalities coexist within the same broader urban areas while concentrating at different junctions. The second analytical module advances towards the quantification of the mechanisms influencing surrogate safety and fuel consumption, at both individual trip and road segment levels. Through the development of two Structural Equation Models (SEM), latent variables capturing driving volatility, road network characteristics, and weather and temporal conditions were identified. Joint modeling results further confirm that high fuel consumption consistently co-occurs with aggressive driving, characterized by frequent harsh acceleration and braking events, and distracted driving associated with mobile phone use across the examined levels, reflecting shared behavioral and contextual mechanisms. In contrast, the relationship between speeding and fuel consumption follows distinct structural pathways that depend primarily on road external context. In the third analytical module, an explainable multi-criteria efficiency analysis was developed to evaluate trip-level driving efficiency by simultaneously considering road safety, fuel consumption, and travel time, and to aggregate these outcomes into road-level efficiency. Results showed that efficient trips are consistently characterized by lower fuel consumption, shorter travel time per kilometer, as well as fewer harsh braking and mobile phone use events. This module showed that the three considered sustainable mobility pillars can be coherently integrated into a single sustainability efficiency metric, while traffic congestion constitutes the main external driver of driving inefficiency in terms of sustainability. Overall, the data-driven findings demonstrate that the relationship between safe and green mobility is significant and characterized by both synergies and trade-offs across levels of analysis. In parallel, they highlight the importance of multilevel and integrated assessment approaches for addressing the multidimensional nature of sustainability and for supporting informed and targeted interventions towards safer and greener mobility for all.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Petraki, Virginia
- Πετράκη, Βιργινία
Subjects
dc:subject × 18- Βιώσιμη κινητικότητα
- Οδική ασφάλεια
- Οδηγική συμπεριφορά
- Ενεργειακή αποδοτικότητα
- Μηχανική μάθηση
- Χωρική ανάλυση
- Sustainable mobility
- Road safety
- Driving behavior
- Energy efficiency
- Machine learning
- Spatial analysis
- Επιστήμη Πολιτικού Μηχανικού
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη και Τεχνολογία των μεταφορών
- Civil Engineering
- Engineering and Technology
- Transportation science and Technology
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/61747