
The Safety Tolerance Zone (STZ) refers to the real-world phenomenon of (technology assisted) drivers self-regulating control over transportation vehicles in the context of crash avoidance. The concept of the STZ attempted to describe the point at which self-regulated control is considered safe. It is the zone where the demands of the driving task are balanced with the ability of the driver to cope with them. The aim of this study was to assess road, vehicle and behavioural risk indicators for the identification of Safety Tolerance Zone (STZ). Towards that end, a simulator driving experiment was carried out involving 55 drivers and a database consisting of 165 trips was created. Participants were asked to complete a driving behaviour questionnaire to collect detailed information on various aspects of driving, socio-demographics, safety attitude and psychological factors. True STZ categories are identified using a variable time-headway threshold, adjusted by task complexity and coping capacity, with fixed boundaries: >2 sec (normal), 1.4-2 sec (danger), and <1.4 sec (avoidable accident).According to the feature importance analysis, time-to-collision, average speed, driving duration, hands-on event and fatigue found to be the most influential factors among all examined indicators. Conversely, parameters such as lane departure warning or forward collision warning had a negligible impact on STZ headway. Findings from the current research highlighted the models’ reliability and effectiveness in classifying instances of headway. The models performed best in normal driving phases, likely because these conditions were more consistent and made up the majority of the training data. Additionally, normal driving features were more distinct and less ambiguous compared to hazardous conditions, reducing misclassification risks. These results not only showcased their potential for real-world applications but also emphasised their significance in the realm of road safety and traffic management.
ID | pc580 |
Presentation | |
Full Text | |
Tags |