Implementation of a simulated environment for data acquisition to monitor drivers' stress levels

Abstract

Most of the research on transportation safety aims to reduce the damage of accidents at the moment they occur or after they happen. Passive safety systems help to reduce the effects of the accidents, while active safety systems serve to minimize the risk of collision. The majority of accidents are related to the mental workload of the driver. In low levels of workload , fatigue and, in high levels of it, stress lead to driver errors which are the number one reason for accidents. In such cases the driver who can still perform routine driving tasks without problems is unable to cope with unattended or extraordinary driving conditions. In literature, there are several studies that propose monitoring of drivers physical and cognitive status and helping them by alerts or stimuli to avoid accidents in critical conditions. Nevertheless, the selection of objective parameters from these measurements is an ongoing problem. Another consideration should be the unobtrusiveness and ease of use for such systems. This study targets a simulated driving environment based on UNITY game engine, where a driver's status can be monitored multimodally. It is shown that a variety of physiological parameters such as changes in skin resistance and heart rate can be recorded under adjustable difficulty driving conditions. This provides a flexible toolkit which enables the analysis of the relation between the driver's stress level and his/her driving ability from different perspectives and without being physically on the road traffic.