Sekadakis M., Trösterer S., Moertl P., Yannis G., “Examining Factors Affecting Takeover Time in Automated Driving using Machine Learning Algorithms”, Proceedings of the 30th Intelligent Transportation Systems (ITS) World Congress, Dubai, 16-20 September 2024

This study, performed under the HADRIAN project, aims to investigate the factors affecting Takeover Time in Automated Driving, a critical safety Key Performance Indicator (KPI). The data were gathered through a driving simulator experiment as well as questionnaires which evaluated innovative HMI prototypes designed to enhance Automated Driving (AD) at SAE Levels 2 and 3. A safety and impact assessment methodology tailored to HADRIAN was developed, incorporating unique KPIs. Descriptive insights for these KPIs were formed to facilitate comparisons and substantiate the observed safety and impact advancements. Moreover, XGBoost feature analysis algorithms were deployed to obtain the most significant KPIs that affect Takeover Time from automated to manual driving. The key influential factors include the driver’s trust in the HMI, followed by the frequency of takeover requests, the smoothness during takeover manoeuvres, and the specific HMI condition (baseline or HADRIAN HMI).