Spyros Tsigkos, “Correlation of crash risk with driver capacity in coping with driving tasks complexity using machine learning”, Diploma Thesis, NTUA, School of Civil Engineering, Athens, July 2023.

The aim of this thesis is to correlate the driver’s ability to cope with the complexity of driving with crash risk using machine learning techniques. Data required for this Diploma Thesis, refer to driving data under real-life conditions in Great Britain, collected in the context of the i-DREAMS research project. To achieve this objective, a descriptive statistical analysis of the database was conducted and eight Structural Equation Models (SEM) were developed. It emerged that increasing the complexity of the trip increases the risk of an accident, while worsening the condition of the driver and the vehicle also increase the risk of an accident. Male drivers have more high-severity abrupt incidents while driving compared to female drivers, which confirms the international literature.