Marios Sekadakis, “Safety Assessment and Behaviour Modelling of Take-Over Dynamics in Automated Driving: From Human Response to Network-Level Simulation”, PhD Thesis, School of Civil Engineering, National Technical University of Athens (November 2021 – January 2026).

This PhD developed an integrated framework to explain how take-over response, driving performance and safety in automated driving at SAE Levels 2 and 3 are shaped by transition context, driver state, interface design and vehicle state dynamics. The work combined evidence synthesis using meta-analysis with random effects meta-regression, controlled driving simulator experimentation, behavioural efficiency assessment through Data Envelopment Analysis (DEA), unsupervised behavioural profiling using k-means clustering, data-driven modelling with XGBoost and SHAP, and spatial Generalised Additive Models (sGAM) applied to calibrated microscopic traffic simulation. This methodological chain enabled the systematic linkage of individual re-engagement behaviour during the transition from automated to manual control with vehicle dynamics, interaction patterns and network-level safety effects. This dissertation demonstrates that safety during take-over transitions cannot be explained by take-over duration alone. By integrating Human-Machine Interface (HMI) design, automation level, behavioural variability and roadway context across complementary analytical methods, it identifies how and where temporal safety margins deteriorate during re-engagement. The findings provide a clear and evidence-based foundation for safer take-over management and more robust evaluation of automated driving systems in mixed traffic.